diff --git a/.env.example b/.env.example
new file mode 100644
index 00000000..b1ceee95
--- /dev/null
+++ b/.env.example
@@ -0,0 +1,3 @@
+PORT=3000
+DEBUG=false
+HTTPS_PROXY=http://localhost:7890
\ No newline at end of file
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
new file mode 100644
index 00000000..3034a547
--- /dev/null
+++ b/.github/workflows/ci.yml
@@ -0,0 +1,47 @@
+name: CI
+
+# This setup assumes that you run the unit tests with code coverage in the same
+# workflow that will also print the coverage report as comment to the pull request.
+# Therefore, you need to trigger this workflow when a pull request is (re)opened or
+# when new code is pushed to the branch of the pull request. In addition, you also
+# need to trigger this workflow when new code is pushed to the main branch because
+# we need to upload the code coverage results as artifact for the main branch as
+# well since it will be the baseline code coverage.
+#
+# We do not want to trigger the workflow for pushes to *any* branch because this
+# would trigger our jobs twice on pull requests (once from "push" event and once
+# from "pull_request->synchronize")
+on:
+  pull_request:
+    types: [opened, reopened, synchronize]
+  push:
+    branches:
+      - 'main'
+
+jobs:
+  unit_tests:
+    name: "Unit tests"
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Setup Go
+        uses: actions/setup-go@v4
+        with:
+          go-version: ^1.22
+
+      # When you execute your unit tests, make sure to use the "-coverprofile" flag to write a
+      # coverage profile to a file. You will need the name of the file (e.g. "coverage.txt")
+      # in the next step as well as the next job.
+      - name: Test
+        run: go test -cover -coverprofile=coverage.txt ./...
+      - uses: codecov/codecov-action@v4
+        with:
+          token: ${{ secrets.CODECOV_TOKEN }}
+
+  commit_lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - uses: wagoid/commitlint-github-action@v6
diff --git a/.github/workflows/docker-image-amd64.yml b/.github/workflows/docker-image-amd64.yml
deleted file mode 100644
index 2079d31f..00000000
--- a/.github/workflows/docker-image-amd64.yml
+++ /dev/null
@@ -1,61 +0,0 @@
-name: Publish Docker image (amd64)
-
-on:
-  push:
-    tags:
-      - '*'
-  workflow_dispatch:
-    inputs:
-      name:
-        description: 'reason'
-        required: false
-jobs:
-  push_to_registries:
-    name: Push Docker image to multiple registries
-    runs-on: ubuntu-latest
-    permissions:
-      packages: write
-      contents: read
-    steps:
-      - name: Check out the repo
-        uses: actions/checkout@v3
-
-      - name: Check repository URL
-        run: |
-          REPO_URL=$(git config --get remote.origin.url)
-          if [[ $REPO_URL == *"pro" ]]; then
-            exit 1
-          fi        
-
-      - name: Save version info
-        run: |
-          git describe --tags > VERSION 
-
-      - name: Log in to Docker Hub
-        uses: docker/login-action@v2
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      - name: Log in to the Container registry
-        uses: docker/login-action@v2
-        with:
-          registry: ghcr.io
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@v4
-        with:
-          images: |
-            justsong/one-api
-            ghcr.io/${{ github.repository }}
-
-      - name: Build and push Docker images
-        uses: docker/build-push-action@v3
-        with:
-          context: .
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
\ No newline at end of file
diff --git a/.github/workflows/docker-image-amd64-en.yml b/.github/workflows/docker-image-en.yml
similarity index 84%
rename from .github/workflows/docker-image-amd64-en.yml
rename to .github/workflows/docker-image-en.yml
index af488256..30cd0e38 100644
--- a/.github/workflows/docker-image-amd64-en.yml
+++ b/.github/workflows/docker-image-en.yml
@@ -1,9 +1,9 @@
-name: Publish Docker image (amd64, English)
+name: Publish Docker image (English)
 
 on:
   push:
     tags:
-      - '*'
+      - 'v*.*.*'
   workflow_dispatch:
     inputs:
       name:
@@ -34,6 +34,13 @@ jobs:
       - name: Translate
         run: |
           python ./i18n/translate.py --repository_path . --json_file_path ./i18n/en.json
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+
       - name: Log in to Docker Hub
         uses: docker/login-action@v2
         with:
@@ -51,6 +58,7 @@ jobs:
         uses: docker/build-push-action@v3
         with:
           context: .
+          platforms: linux/amd64,linux/arm64
           push: true
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
\ No newline at end of file
diff --git a/.github/workflows/docker-image-arm64.yml b/.github/workflows/docker-image.yml
similarity index 96%
rename from .github/workflows/docker-image-arm64.yml
rename to .github/workflows/docker-image.yml
index 39d1a401..56f1d6ad 100644
--- a/.github/workflows/docker-image-arm64.yml
+++ b/.github/workflows/docker-image.yml
@@ -1,10 +1,9 @@
-name: Publish Docker image (arm64)
+name: Publish Docker image
 
 on:
   push:
     tags:
-      - '*'
-      - '!*-alpha*'
+      - 'v*.*.*'
   workflow_dispatch:
     inputs:
       name:
diff --git a/.github/workflows/linux-release.yml b/.github/workflows/linux-release.yml
index 6f30a1d5..161c41e3 100644
--- a/.github/workflows/linux-release.yml
+++ b/.github/workflows/linux-release.yml
@@ -5,7 +5,7 @@ permissions:
 on:
   push:
     tags:
-      - '*'
+      - 'v*.*.*'
       - '!*-alpha*'
   workflow_dispatch:
     inputs:
diff --git a/.github/workflows/macos-release.yml b/.github/workflows/macos-release.yml
index 359c2c92..94b3e47b 100644
--- a/.github/workflows/macos-release.yml
+++ b/.github/workflows/macos-release.yml
@@ -5,7 +5,7 @@ permissions:
 on:
   push:
     tags:
-      - '*'
+      - 'v*.*.*'
       - '!*-alpha*'
   workflow_dispatch:
     inputs:
diff --git a/.github/workflows/windows-release.yml b/.github/workflows/windows-release.yml
index 4e99b75c..18641ae8 100644
--- a/.github/workflows/windows-release.yml
+++ b/.github/workflows/windows-release.yml
@@ -5,7 +5,7 @@ permissions:
 on:
   push:
     tags:
-      - '*'
+      - 'v*.*.*'
       - '!*-alpha*'
   workflow_dispatch:
     inputs:
diff --git a/.gitignore b/.gitignore
index 974fcf63..0cedb4b4 100644
--- a/.gitignore
+++ b/.gitignore
@@ -8,3 +8,6 @@ build
 logs
 data
 /web/node_modules
+cmd.md
+.env
+/one-api
diff --git a/Dockerfile b/Dockerfile
index 10301631..7902b37c 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,4 +1,4 @@
-FROM node:16 as builder
+FROM --platform=$BUILDPLATFORM node:16 AS builder
 
 WORKDIR /web
 COPY ./VERSION .
@@ -16,7 +16,9 @@ WORKDIR /web/air
 RUN npm install --force
 RUN DISABLE_ESLINT_PLUGIN='true' REACT_APP_VERSION=$(cat VERSION) npm run build
 
-FROM golang AS builder2
+FROM golang:alpine AS builder2
+
+RUN apk add --no-cache g++
 
 ENV GO111MODULE=on \
     CGO_ENABLED=1 \
@@ -27,7 +29,7 @@ ADD go.mod go.sum ./
 RUN go mod download
 COPY . .
 COPY --from=builder /web/build ./web/build
-RUN go build -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
+RUN go build -trimpath -ldflags "-s -w -X 'github.com/songquanpeng/one-api/common.Version=$(cat VERSION)' -extldflags '-static'" -o one-api
 
 FROM alpine
 
diff --git a/README.en.md b/README.en.md
index bce47353..c9fdbbc8 100644
--- a/README.en.md
+++ b/README.en.md
@@ -101,7 +101,7 @@ Nginx reference configuration:
 ```
 server{
    server_name openai.justsong.cn;  # Modify your domain name accordingly
-   
+
    location / {
           client_max_body_size  64m;
           proxy_http_version 1.1;
@@ -132,12 +132,12 @@ The initial account username is `root` and password is `123456`.
 1. Download the executable file from [GitHub Releases](https://github.com/songquanpeng/one-api/releases/latest) or compile from source:
    ```shell
    git clone https://github.com/songquanpeng/one-api.git
-   
+
    # Build the frontend
    cd one-api/web/default
    npm install
    npm run build
-   
+
    # Build the backend
    cd ../..
    go mod download
@@ -245,16 +245,41 @@ If the channel ID is not provided, load balancing will be used to distribute the
     + Example: `LOG_SQL_DSN=root:123456@tcp(localhost:3306)/oneapi-logs`
 5. `FRONTEND_BASE_URL`: When set, the specified frontend address will be used instead of the backend address.
     + Example: `FRONTEND_BASE_URL=https://openai.justsong.cn`
-6. `SYNC_FREQUENCY`: When set, the system will periodically sync configurations from the database, with the unit in seconds. If not set, no sync will happen.
+6. 'MEMORY_CACHE_ENABLED': Enabling memory caching can cause a certain delay in updating user quotas, with optional values of 'true' and 'false'. If not set, it defaults to 'false'.
+7. `SYNC_FREQUENCY`: When set, the system will periodically sync configurations from the database, with the unit in seconds. If not set, no sync will happen.
     + Example: `SYNC_FREQUENCY=60`
-7. `NODE_TYPE`: When set, specifies the node type. Valid values are `master` and `slave`. If not set, it defaults to `master`.
+8. `NODE_TYPE`: When set, specifies the node type. Valid values are `master` and `slave`. If not set, it defaults to `master`.
     + Example: `NODE_TYPE=slave`
-8. `CHANNEL_UPDATE_FREQUENCY`: When set, it periodically updates the channel balances, with the unit in minutes. If not set, no update will happen.
+9. `CHANNEL_UPDATE_FREQUENCY`: When set, it periodically updates the channel balances, with the unit in minutes. If not set, no update will happen.
     + Example: `CHANNEL_UPDATE_FREQUENCY=1440`
-9. `CHANNEL_TEST_FREQUENCY`: When set, it periodically tests the channels, with the unit in minutes. If not set, no test will happen.
+10. `CHANNEL_TEST_FREQUENCY`: When set, it periodically tests the channels, with the unit in minutes. If not set, no test will happen.
     + Example: `CHANNEL_TEST_FREQUENCY=1440`
-10. `POLLING_INTERVAL`: The time interval (in seconds) between requests when updating channel balances and testing channel availability. Default is no interval.
+11. `POLLING_INTERVAL`: The time interval (in seconds) between requests when updating channel balances and testing channel availability. Default is no interval.
     + Example: `POLLING_INTERVAL=5`
+12. `BATCH_UPDATE_ENABLED`: Enabling batch database update aggregation can cause a certain delay in updating user quotas. The optional values are 'true' and 'false', but if not set, it defaults to 'false'.
+    +Example: ` BATCH_UPDATE_ENABLED=true`
+    +If you encounter an issue with too many database connections, you can try enabling this option.
+13. `BATCH_UPDATE_INTERVAL=5`: The time interval for batch updating aggregates, measured in seconds, defaults to '5'.
+    +Example: ` BATCH_UPDATE_INTERVAL=5`
+14. Request frequency limit:
+    + `GLOBAL_API_RATE_LIMIT`: Global API rate limit (excluding relay requests), the maximum number of requests within three minutes per IP, default to 180.
+    + `GLOBAL_WEL_RATE_LIMIT`: Global web speed limit, the maximum number of requests within three minutes per IP, default to 60.
+15. Encoder cache settings:
+    +`TIKTOKEN_CACHE_DIR`: By default, when the program starts, it will download the encoding of some common word elements online, such as' gpt-3.5 turbo '. In some unstable network environments or offline situations, it may cause startup problems. This directory can be configured to cache data and can be migrated to an offline environment.
+    +`DATA_GYM_CACHE_DIR`: Currently, this configuration has the same function as' TIKTOKEN-CACHE-DIR ', but its priority is not as high as it.
+16. `RELAY_TIMEOUT`: Relay timeout setting, measured in seconds, with no default timeout time set.
+17. `RELAY_PROXY`: After setting up, use this proxy to request APIs.
+18. `USER_CONTENT_REQUEST_TIMEOUT`: The timeout period for users to upload and download content, measured in seconds.
+19. `USER_CONTENT_REQUEST_PROXY`: After setting up, use this agent to request content uploaded by users, such as images.
+20. `SQLITE_BUSY_TIMEOUT`: SQLite lock wait timeout setting, measured in milliseconds, default to '3000'.
+21. `GEMINI_SAFETY_SETTING`: Gemini's security settings are set to 'BLOCK-NONE' by default.
+22. `GEMINI_VERSION`: The Gemini version used by the One API, which defaults to 'v1'.
+23. `THE`: The system's theme setting, default to 'default', specific optional values refer to [here] (./web/README. md).
+24. `ENABLE_METRIC`: Whether to disable channels based on request success rate, default not enabled, optional values are 'true' and 'false'.
+25. `METRIC_QUEUE_SIZE`: Request success rate statistics queue size, default to '10'.
+26. `METRIC_SUCCESS_RATE_THRESHOLD`: Request success rate threshold, default to '0.8'.
+27. `INITIAL_ROOT_TOKEN`: If this value is set, a root user token with the value of the environment variable will be automatically created when the system starts for the first time.
+28. `INITIAL_ROOT_ACCESS_TOKEN`: If this value is set, a system management token will be automatically created for the root user with a value of the environment variable when the system starts for the first time.
 
 ### Command Line Parameters
 1. `--port <port_number>`: Specifies the port number on which the server listens. Defaults to `3000`.
@@ -287,7 +312,9 @@ If the channel ID is not provided, load balancing will be used to distribute the
     + Double-check that your interface address and API Key are correct.
 
 ## Related Projects
-[FastGPT](https://github.com/labring/FastGPT): Knowledge question answering system based on the LLM
+* [FastGPT](https://github.com/labring/FastGPT): Knowledge question answering system based on the LLM
+* [VChart](https://github.com/VisActor/VChart):  More than just a cross-platform charting library, but also an expressive data storyteller.
+* [VMind](https://github.com/VisActor/VMind):  Not just automatic, but also fantastic. Open-source solution for intelligent visualization.
 
 ## Note
 This project is an open-source project. Please use it in compliance with OpenAI's [Terms of Use](https://openai.com/policies/terms-of-use) and **applicable laws and regulations**. It must not be used for illegal purposes.
diff --git a/README.md b/README.md
index efb04b6c..853ec067 100644
--- a/README.md
+++ b/README.md
@@ -53,7 +53,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用 
 
 > [!NOTE]
 > 本项目为开源项目，使用者必须在遵循 OpenAI 的[使用条款](https://openai.com/policies/terms-of-use)以及**法律法规**的情况下使用，不得用于非法用途。
-> 
+>
 > 根据[《生成式人工智能服务管理暂行办法》](http://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm)的要求，请勿对中国地区公众提供一切未经备案的生成式人工智能服务。
 
 > [!WARNING]
@@ -65,9 +65,10 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用 
 ## 功能
 1. 支持多种大模型：
    + [x] [OpenAI ChatGPT 系列模型](https://platform.openai.com/docs/guides/gpt/chat-completions-api)（支持 [Azure OpenAI API](https://learn.microsoft.com/en-us/azure/ai-services/openai/reference)）
-   + [x] [Anthropic Claude 系列模型](https://anthropic.com)
+   + [x] [Anthropic Claude 系列模型](https://anthropic.com) (支持 AWS Claude)
    + [x] [Google PaLM2/Gemini 系列模型](https://developers.generativeai.google)
    + [x] [Mistral 系列模型](https://mistral.ai/)
+   + [x] [字节跳动豆包大模型](https://console.volcengine.com/ark/region:ark+cn-beijing/model)
    + [x] [百度文心一言系列模型](https://cloud.baidu.com/doc/WENXINWORKSHOP/index.html)
    + [x] [阿里通义千问系列模型](https://help.aliyun.com/document_detail/2400395.html)
    + [x] [讯飞星火认知大模型](https://www.xfyun.cn/doc/spark/Web.html)
@@ -76,18 +77,27 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用 
    + [x] [腾讯混元大模型](https://cloud.tencent.com/document/product/1729)
    + [x] [Moonshot AI](https://platform.moonshot.cn/)
    + [x] [百川大模型](https://platform.baichuan-ai.com)
-   + [ ] [字节云雀大模型](https://www.volcengine.com/product/ark) (WIP)
    + [x] [MINIMAX](https://api.minimax.chat/)
    + [x] [Groq](https://wow.groq.com/)
    + [x] [Ollama](https://github.com/ollama/ollama)
    + [x] [零一万物](https://platform.lingyiwanwu.com/)
+   + [x] [阶跃星辰](https://platform.stepfun.com/)
+   + [x] [Coze](https://www.coze.com/)
+   + [x] [Cohere](https://cohere.com/)
+   + [x] [DeepSeek](https://www.deepseek.com/)
+   + [x] [Cloudflare Workers AI](https://developers.cloudflare.com/workers-ai/)
+   + [x] [DeepL](https://www.deepl.com/)
+   + [x] [together.ai](https://www.together.ai/)
+   + [x] [novita.ai](https://www.novita.ai/)
+   + [x] [硅基流动 SiliconCloud](https://siliconflow.cn/siliconcloud)
+   + [x] [xAI](https://x.ai/)
 2. 支持配置镜像以及众多[第三方代理服务](https://iamazing.cn/page/openai-api-third-party-services)。
 3. 支持通过**负载均衡**的方式访问多个渠道。
 4. 支持 **stream 模式**，可以通过流式传输实现打字机效果。
 5. 支持**多机部署**，[详见此处](#多机部署)。
-6. 支持**令牌管理**，设置令牌的过期时间和额度。
+6. 支持**令牌管理**，设置令牌的过期时间、额度、允许的 IP 范围以及允许的模型访问。
 7. 支持**兑换码管理**，支持批量生成和导出兑换码，可使用兑换码为账户进行充值。
-8. 支持**通道管理**，批量创建通道。
+8. 支持**渠道管理**，批量创建渠道。
 9. 支持**用户分组**以及**渠道分组**，支持为不同分组设置不同的倍率。
 10. 支持渠道**设置模型列表**。
 11. 支持**查看额度明细**。
@@ -101,11 +111,12 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用 
 19. 支持丰富的**自定义**设置，
     1. 支持自定义系统名称，logo 以及页脚。
     2. 支持自定义首页和关于页面，可以选择使用 HTML & Markdown 代码进行自定义，或者使用一个单独的网页通过 iframe 嵌入。
-20. 支持通过系统访问令牌访问管理 API（bearer token，用以替代 cookie，你可以自行抓包来查看 API 的用法）。
+20. 支持通过系统访问令牌调用管理 API，进而**在无需二开的情况下扩展和自定义** One API 的功能，详情请参考此处 [API 文档](./docs/API.md)。。
 21. 支持 Cloudflare Turnstile 用户校验。
 22. 支持用户管理，支持**多种用户登录注册方式**：
     + 邮箱登录注册（支持注册邮箱白名单）以及通过邮箱进行密码重置。
-    + [GitHub 开放授权](https://github.com/settings/applications/new)。
+    + 支持[飞书授权登录](https://open.feishu.cn/document/uAjLw4CM/ukTMukTMukTM/reference/authen-v1/authorize/get)（[这里有 One API 的实现细节阐述供参考](https://iamazing.cn/page/feishu-oauth-login)）。
+    + 支持 [GitHub 授权登录](https://github.com/settings/applications/new)。
     + 微信公众号授权（需要额外部署 [WeChat Server](https://github.com/songquanpeng/wechat-server)）。
 23. 支持主题切换，设置环境变量 `THEME` 即可，默认为 `default`，欢迎 PR 更多主题，具体参考[此处](./web/README.md)。
 24. 配合 [Message Pusher](https://github.com/songquanpeng/message-pusher) 可将报警信息推送到多种 App 上。
@@ -136,7 +147,7 @@ Nginx 的参考配置：
 ```
 server{
    server_name openai.justsong.cn;  # 请根据实际情况修改你的域名
-   
+
    location / {
           client_max_body_size  64m;
           proxy_http_version 1.1;
@@ -164,6 +175,10 @@ sudo service nginx restart
 
 初始账号用户名为 `root`，密码为 `123456`。
 
+### 通过宝塔面板进行一键部署
+1. 安装宝塔面板9.2.0及以上版本，前往 [宝塔面板](https://www.bt.cn/new/download.html?r=dk_oneapi) 官网，选择正式版的脚本下载安装；
+2. 安装后登录宝塔面板，在左侧菜单栏中点击 `Docker`，首次进入会提示安装 `Docker` 服务，点击立即安装，按提示完成安装；
+3. 安装完成后在应用商店中搜索 `One-API`，点击安装，配置域名等基本信息即可完成安装；
 
 ### 基于 Docker Compose 进行部署
 
@@ -181,12 +196,12 @@ docker-compose ps
 1. 从 [GitHub Releases](https://github.com/songquanpeng/one-api/releases/latest) 下载可执行文件或者从源码编译：
    ```shell
    git clone https://github.com/songquanpeng/one-api.git
-   
+
    # 构建前端
    cd one-api/web/default
    npm install
    npm run build
-   
+
    # 构建后端
    cd ../..
    go mod download
@@ -207,7 +222,7 @@ docker-compose ps
 3. 所有从服务器必须设置 `NODE_TYPE` 为 `slave`，不设置则默认为主服务器。
 4. 设置 `SYNC_FREQUENCY` 后服务器将定期从数据库同步配置，在使用远程数据库的情况下，推荐设置该项并启用 Redis，无论主从。
 5. 从服务器可以选择设置 `FRONTEND_BASE_URL`，以重定向页面请求到主服务器。
-6. 从服务器上**分别**装好 Redis，设置好 `REDIS_CONN_STRING`，这样可以做到在缓存未过期的情况下数据库零访问，可以减少延迟。
+6. 从服务器上**分别**装好 Redis，设置好 `REDIS_CONN_STRING`，这样可以做到在缓存未过期的情况下数据库零访问，可以减少延迟（Redis 集群或者哨兵模式的支持请参考环境变量说明）。
 7. 如果主服务器访问数据库延迟也比较高，则也需要启用 Redis，并设置 `SYNC_FREQUENCY`，以定期从数据库同步配置。
 
 环境变量的具体使用方法详见[此处](#环境变量)。
@@ -242,9 +257,9 @@ docker run --name chatgpt-web -d -p 3002:3002 -e OPENAI_API_BASE_URL=https://ope
 #### QChatGPT - QQ机器人
 项目主页：https://github.com/RockChinQ/QChatGPT
 
-根据文档完成部署后，在`config.py`设置配置项`openai_config`的`reverse_proxy`为 One API 后端地址，设置`api_key`为 One API 生成的key，并在配置项`completion_api_params`的`model`参数设置为 One API 支持的模型名称。
+根据[文档](https://qchatgpt.rockchin.top)完成部署后，在 `data/provider.json`设置`requester.openai-chat-completions.base-url`为 One API 实例地址，并填写 API Key 到 `keys.openai` 组中，设置 `model` 为要使用的模型名称。
 
-可安装 [Switcher 插件](https://github.com/RockChinQ/Switcher)在运行时切换所使用的模型。
+运行期间可以通过`!model`命令查看、切换可用模型。
 
 ### 部署到第三方平台
 <details>
@@ -313,7 +328,7 @@ Render 可以直接部署 docker 镜像，不需要 fork 仓库：https://dashbo
 例如对于 OpenAI 的官方库：
 ```bash
 OPENAI_API_KEY="sk-xxxxxx"
-OPENAI_API_BASE="https://<HOST>:<PORT>/v1" 
+OPENAI_API_BASE="https://<HOST>:<PORT>/v1"
 ```
 
 ```mermaid
@@ -332,9 +347,15 @@ graph LR
 不加的话将会使用负载均衡的方式使用多个渠道。
 
 ### 环境变量
+> One API 支持从 `.env` 文件中读取环境变量，请参照 `.env.example` 文件，使用时请将其重命名为 `.env`。
 1. `REDIS_CONN_STRING`：设置之后将使用 Redis 作为缓存使用。
    + 例子：`REDIS_CONN_STRING=redis://default:redispw@localhost:49153`
    + 如果数据库访问延迟很低，没有必要启用 Redis，启用后反而会出现数据滞后的问题。
+   + 如果需要使用哨兵或者集群模式：
+     + 则需要把该环境变量设置为节点列表，例如：`localhost:49153,localhost:49154,localhost:49155`。
+     + 除此之外还需要设置以下环境变量：
+       + `REDIS_PASSWORD`：Redis 集群或者哨兵模式下的密码设置。
+       + `REDIS_MASTER_NAME`：Redis 哨兵模式下主节点的名称。
 2. `SESSION_SECRET`：设置之后将使用固定的会话密钥，这样系统重新启动后已登录用户的 cookie 将依旧有效。
    + 例子：`SESSION_SECRET=random_string`
 3. `SQL_DSN`：设置之后将使用指定数据库而非 SQLite，请使用 MySQL 或 PostgreSQL。
@@ -360,8 +381,8 @@ graph LR
    + 例子：`NODE_TYPE=slave`
 9. `CHANNEL_UPDATE_FREQUENCY`：设置之后将定期更新渠道余额，单位为分钟，未设置则不进行更新。
    + 例子：`CHANNEL_UPDATE_FREQUENCY=1440`
-10. `CHANNEL_TEST_FREQUENCY`：设置之后将定期检查渠道，单位为分钟，未设置则不进行检查。
-   + 例子：`CHANNEL_TEST_FREQUENCY=1440`
+10. `CHANNEL_TEST_FREQUENCY`：设置之后将定期检查渠道，单位为分钟，未设置则不进行检查。 
+   +例子：`CHANNEL_TEST_FREQUENCY=1440`
 11. `POLLING_INTERVAL`：批量更新渠道余额以及测试可用性时的请求间隔，单位为秒，默认无间隔。
     + 例子：`POLLING_INTERVAL=5`
 12. `BATCH_UPDATE_ENABLED`：启用数据库批量更新聚合，会导致用户额度的更新存在一定的延迟可选值为 `true` 和 `false`，未设置则默认为 `false`。
@@ -376,13 +397,19 @@ graph LR
     + `TIKTOKEN_CACHE_DIR`：默认程序启动时会联网下载一些通用的词元的编码，如：`gpt-3.5-turbo`，在一些网络环境不稳定，或者离线情况，可能会导致启动有问题，可以配置此目录缓存数据，可迁移到离线环境。
     + `DATA_GYM_CACHE_DIR`：目前该配置作用与 `TIKTOKEN_CACHE_DIR` 一致，但是优先级没有它高。
 16. `RELAY_TIMEOUT`：中继超时设置，单位为秒，默认不设置超时时间。
-17. `SQLITE_BUSY_TIMEOUT`：SQLite 锁等待超时设置，单位为毫秒，默认 `3000`。
-18. `GEMINI_SAFETY_SETTING`：Gemini 的安全设置，默认 `BLOCK_NONE`。
-19. `THEME`：系统的主题设置，默认为 `default`，具体可选值参考[此处](./web/README.md)。
-20. `ENABLE_METRIC`：是否根据请求成功率禁用渠道，默认不开启，可选值为 `true` 和 `false`。
-21. `METRIC_QUEUE_SIZE`：请求成功率统计队列大小，默认为 `10`。
-22. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
-23. `INITIAL_ROOT_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量值的 root 用户令牌。
+17. `RELAY_PROXY`：设置后使用该代理来请求 API。
+18. `USER_CONTENT_REQUEST_TIMEOUT`：用户上传内容下载超时时间，单位为秒。
+19. `USER_CONTENT_REQUEST_PROXY`：设置后使用该代理来请求用户上传的内容，例如图片。
+20. `SQLITE_BUSY_TIMEOUT`：SQLite 锁等待超时设置，单位为毫秒，默认 `3000`。
+21. `GEMINI_SAFETY_SETTING`：Gemini 的安全设置，默认 `BLOCK_NONE`。
+22. `GEMINI_VERSION`：One API 所使用的 Gemini 版本，默认为 `v1`。
+23. `THEME`：系统的主题设置，默认为 `default`，具体可选值参考[此处](./web/README.md)。
+24. `ENABLE_METRIC`：是否根据请求成功率禁用渠道，默认不开启，可选值为 `true` 和 `false`。
+25. `METRIC_QUEUE_SIZE`：请求成功率统计队列大小，默认为 `10`。
+26. `METRIC_SUCCESS_RATE_THRESHOLD`：请求成功率阈值，默认为 `0.8`。
+27. `INITIAL_ROOT_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量值的 root 用户令牌。
+28. `INITIAL_ROOT_ACCESS_TOKEN`：如果设置了该值，则在系统首次启动时会自动创建一个值为该环境变量的 root 用户创建系统管理令牌。
+29. `ENFORCE_INCLUDE_USAGE`：是否强制在 stream 模型下返回 usage，默认不开启，可选值为 `true` 和 `false`。
 
 ### 命令行参数
 1. `--port <port_number>`: 指定服务器监听的端口号，默认为 `3000`。
@@ -421,7 +448,7 @@ https://openai.justsong.cn
    + 检查你的接口地址和 API Key 有没有填对。
    + 检查是否启用了 HTTPS，浏览器会拦截 HTTPS 域名下的 HTTP 请求。
 6. 报错：`当前分组负载已饱和，请稍后再试`
-   + 上游通道 429 了。
+   + 上游渠道 429 了。
 7. 升级之后我的数据会丢失吗？
    + 如果使用 MySQL，不会。
    + 如果使用 SQLite，需要按照我所给的部署命令挂载 volume 持久化 one-api.db 数据库文件，否则容器重启后数据会丢失。
@@ -429,12 +456,14 @@ https://openai.justsong.cn
    + 一般情况下不需要，系统将在初始化的时候自动调整。
    + 如果需要的话，我会在更新日志中说明，并给出脚本。
 9. 手动修改数据库后报错：`数据库一致性已被破坏，请联系管理员`？
-   + 这是检测到 ability 表里有些记录的通道 id 是不存在的，这大概率是因为你删了 channel 表里的记录但是没有同步在 ability 表里清理无效的通道。
-   + 对于每一个通道，其所支持的模型都需要有一个专门的 ability 表的记录，表示该通道支持该模型。
+   + 这是检测到 ability 表里有些记录的渠道 id 是不存在的，这大概率是因为你删了 channel 表里的记录但是没有同步在 ability 表里清理无效的渠道。
+   + 对于每一个渠道，其所支持的模型都需要有一个专门的 ability 表的记录，表示该渠道支持该模型。
 
 ## 相关项目
 * [FastGPT](https://github.com/labring/FastGPT): 基于 LLM 大语言模型的知识库问答系统
 * [ChatGPT Next Web](https://github.com/Yidadaa/ChatGPT-Next-Web):  一键拥有你自己的跨平台 ChatGPT 应用
+* [VChart](https://github.com/VisActor/VChart):  不只是开箱即用的多端图表库，更是生动灵活的数据故事讲述者。
+* [VMind](https://github.com/VisActor/VMind):  不仅自动，还很智能。开源智能可视化解决方案。
 
 ## 注意
 
diff --git a/common/client/init.go b/common/client/init.go
new file mode 100644
index 00000000..f803cbf8
--- /dev/null
+++ b/common/client/init.go
@@ -0,0 +1,60 @@
+package client
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"net/http"
+	"net/url"
+	"time"
+)
+
+var HTTPClient *http.Client
+var ImpatientHTTPClient *http.Client
+var UserContentRequestHTTPClient *http.Client
+
+func Init() {
+	if config.UserContentRequestProxy != "" {
+		logger.SysLog(fmt.Sprintf("using %s as proxy to fetch user content", config.UserContentRequestProxy))
+		proxyURL, err := url.Parse(config.UserContentRequestProxy)
+		if err != nil {
+			logger.FatalLog(fmt.Sprintf("USER_CONTENT_REQUEST_PROXY set but invalid: %s", config.UserContentRequestProxy))
+		}
+		transport := &http.Transport{
+			Proxy: http.ProxyURL(proxyURL),
+		}
+		UserContentRequestHTTPClient = &http.Client{
+			Transport: transport,
+			Timeout:   time.Second * time.Duration(config.UserContentRequestTimeout),
+		}
+	} else {
+		UserContentRequestHTTPClient = &http.Client{}
+	}
+	var transport http.RoundTripper
+	if config.RelayProxy != "" {
+		logger.SysLog(fmt.Sprintf("using %s as api relay proxy", config.RelayProxy))
+		proxyURL, err := url.Parse(config.RelayProxy)
+		if err != nil {
+			logger.FatalLog(fmt.Sprintf("USER_CONTENT_REQUEST_PROXY set but invalid: %s", config.UserContentRequestProxy))
+		}
+		transport = &http.Transport{
+			Proxy: http.ProxyURL(proxyURL),
+		}
+	}
+
+	if config.RelayTimeout == 0 {
+		HTTPClient = &http.Client{
+			Transport: transport,
+		}
+	} else {
+		HTTPClient = &http.Client{
+			Timeout:   time.Duration(config.RelayTimeout) * time.Second,
+			Transport: transport,
+		}
+	}
+
+	ImpatientHTTPClient = &http.Client{
+		Timeout:   5 * time.Second,
+		Transport: transport,
+	}
+}
diff --git a/common/config/config.go b/common/config/config.go
index 3524183a..2eb894ef 100644
--- a/common/config/config.go
+++ b/common/config/config.go
@@ -4,6 +4,7 @@ import (
 	"github.com/songquanpeng/one-api/common/env"
 	"os"
 	"strconv"
+	"strings"
 	"sync"
 	"time"
 
@@ -34,6 +35,7 @@ var PasswordLoginEnabled = true
 var PasswordRegisterEnabled = true
 var EmailVerificationEnabled = false
 var GitHubOAuthEnabled = false
+var OidcEnabled = false
 var WeChatAuthEnabled = false
 var TurnstileCheckEnabled = false
 var RegisterEnabled = true
@@ -51,9 +53,9 @@ var EmailDomainWhitelist = []string{
 	"foxmail.com",
 }
 
-var DebugEnabled = os.Getenv("DEBUG") == "true"
-var DebugSQLEnabled = os.Getenv("DEBUG_SQL") == "true"
-var MemoryCacheEnabled = os.Getenv("MEMORY_CACHE_ENABLED") == "true"
+var DebugEnabled = strings.ToLower(os.Getenv("DEBUG")) == "true"
+var DebugSQLEnabled = strings.ToLower(os.Getenv("DEBUG_SQL")) == "true"
+var MemoryCacheEnabled = strings.ToLower(os.Getenv("MEMORY_CACHE_ENABLED")) == "true"
 
 var LogConsumeEnabled = true
 
@@ -66,6 +68,16 @@ var SMTPToken = ""
 var GitHubClientId = ""
 var GitHubClientSecret = ""
 
+var LarkClientId = ""
+var LarkClientSecret = ""
+
+var OidcClientId = ""
+var OidcClientSecret = ""
+var OidcWellKnown = ""
+var OidcAuthorizationEndpoint = ""
+var OidcTokenEndpoint = ""
+var OidcUserinfoEndpoint = ""
+
 var WeChatServerAddress = ""
 var WeChatServerToken = ""
 var WeChatAccountQRCodeImageURL = ""
@@ -113,10 +125,10 @@ var ValidThemes = map[string]bool{
 // All duration's unit is seconds
 // Shouldn't larger then RateLimitKeyExpirationDuration
 var (
-	GlobalApiRateLimitNum            = env.Int("GLOBAL_API_RATE_LIMIT", 180)
+	GlobalApiRateLimitNum            = env.Int("GLOBAL_API_RATE_LIMIT", 240)
 	GlobalApiRateLimitDuration int64 = 3 * 60
 
-	GlobalWebRateLimitNum            = env.Int("GLOBAL_WEB_RATE_LIMIT", 60)
+	GlobalWebRateLimitNum            = env.Int("GLOBAL_WEB_RATE_LIMIT", 120)
 	GlobalWebRateLimitDuration int64 = 3 * 60
 
 	UploadRateLimitNum            = 10
@@ -138,3 +150,15 @@ var MetricSuccessChanSize = env.Int("METRIC_SUCCESS_CHAN_SIZE", 1024)
 var MetricFailChanSize = env.Int("METRIC_FAIL_CHAN_SIZE", 128)
 
 var InitialRootToken = os.Getenv("INITIAL_ROOT_TOKEN")
+
+var InitialRootAccessToken = os.Getenv("INITIAL_ROOT_ACCESS_TOKEN")
+
+var GeminiVersion = env.String("GEMINI_VERSION", "v1")
+
+var OnlyOneLogFile = env.Bool("ONLY_ONE_LOG_FILE", false)
+
+var RelayProxy = env.String("RELAY_PROXY", "")
+var UserContentRequestProxy = env.String("USER_CONTENT_REQUEST_PROXY", "")
+var UserContentRequestTimeout = env.Int("USER_CONTENT_REQUEST_TIMEOUT", 30)
+
+var EnforceIncludeUsage = env.Bool("ENFORCE_INCLUDE_USAGE", false)
diff --git a/common/constants.go b/common/constants.go
index 849bdce7..87221b61 100644
--- a/common/constants.go
+++ b/common/constants.go
@@ -4,116 +4,3 @@ import "time"
 
 var StartTime = time.Now().Unix() // unit: second
 var Version = "v0.0.0"            // this hard coding will be replaced automatically when building, no need to manually change
-
-const (
-	RoleGuestUser  = 0
-	RoleCommonUser = 1
-	RoleAdminUser  = 10
-	RoleRootUser   = 100
-)
-
-const (
-	UserStatusEnabled  = 1 // don't use 0, 0 is the default value!
-	UserStatusDisabled = 2 // also don't use 0
-	UserStatusDeleted  = 3
-)
-
-const (
-	TokenStatusEnabled   = 1 // don't use 0, 0 is the default value!
-	TokenStatusDisabled  = 2 // also don't use 0
-	TokenStatusExpired   = 3
-	TokenStatusExhausted = 4
-)
-
-const (
-	RedemptionCodeStatusEnabled  = 1 // don't use 0, 0 is the default value!
-	RedemptionCodeStatusDisabled = 2 // also don't use 0
-	RedemptionCodeStatusUsed     = 3 // also don't use 0
-)
-
-const (
-	ChannelStatusUnknown          = 0
-	ChannelStatusEnabled          = 1 // don't use 0, 0 is the default value!
-	ChannelStatusManuallyDisabled = 2 // also don't use 0
-	ChannelStatusAutoDisabled     = 3
-)
-
-const (
-	ChannelTypeUnknown = iota
-	ChannelTypeOpenAI
-	ChannelTypeAPI2D
-	ChannelTypeAzure
-	ChannelTypeCloseAI
-	ChannelTypeOpenAISB
-	ChannelTypeOpenAIMax
-	ChannelTypeOhMyGPT
-	ChannelTypeCustom
-	ChannelTypeAILS
-	ChannelTypeAIProxy
-	ChannelTypePaLM
-	ChannelTypeAPI2GPT
-	ChannelTypeAIGC2D
-	ChannelTypeAnthropic
-	ChannelTypeBaidu
-	ChannelTypeZhipu
-	ChannelTypeAli
-	ChannelTypeXunfei
-	ChannelType360
-	ChannelTypeOpenRouter
-	ChannelTypeAIProxyLibrary
-	ChannelTypeFastGPT
-	ChannelTypeTencent
-	ChannelTypeGemini
-	ChannelTypeMoonshot
-	ChannelTypeBaichuan
-	ChannelTypeMinimax
-	ChannelTypeMistral
-	ChannelTypeGroq
-	ChannelTypeOllama
-	ChannelTypeLingYiWanWu
-
-	ChannelTypeDummy
-)
-
-var ChannelBaseURLs = []string{
-	"",                              // 0
-	"https://api.openai.com",        // 1
-	"https://oa.api2d.net",          // 2
-	"",                              // 3
-	"https://api.closeai-proxy.xyz", // 4
-	"https://api.openai-sb.com",     // 5
-	"https://api.openaimax.com",     // 6
-	"https://api.ohmygpt.com",       // 7
-	"",                              // 8
-	"https://api.caipacity.com",     // 9
-	"https://api.aiproxy.io",        // 10
-	"https://generativelanguage.googleapis.com", // 11
-	"https://api.api2gpt.com",                   // 12
-	"https://api.aigc2d.com",                    // 13
-	"https://api.anthropic.com",                 // 14
-	"https://aip.baidubce.com",                  // 15
-	"https://open.bigmodel.cn",                  // 16
-	"https://dashscope.aliyuncs.com",            // 17
-	"",                                          // 18
-	"https://ai.360.cn",                         // 19
-	"https://openrouter.ai/api",                 // 20
-	"https://api.aiproxy.io",                    // 21
-	"https://fastgpt.run/api/openapi",           // 22
-	"https://hunyuan.cloud.tencent.com",         // 23
-	"https://generativelanguage.googleapis.com", // 24
-	"https://api.moonshot.cn",                   // 25
-	"https://api.baichuan-ai.com",               // 26
-	"https://api.minimax.chat",                  // 27
-	"https://api.mistral.ai",                    // 28
-	"https://api.groq.com/openai",               // 29
-	"http://localhost:11434",                    // 30
-	"https://api.lingyiwanwu.com",               // 31
-}
-
-const (
-	ConfigKeyPrefix = "cfg_"
-
-	ConfigKeyAPIVersion = ConfigKeyPrefix + "api_version"
-	ConfigKeyLibraryID  = ConfigKeyPrefix + "library_id"
-	ConfigKeyPlugin     = ConfigKeyPrefix + "plugin"
-)
diff --git a/common/conv/any.go b/common/conv/any.go
new file mode 100644
index 00000000..467e8bb7
--- /dev/null
+++ b/common/conv/any.go
@@ -0,0 +1,6 @@
+package conv
+
+func AsString(v any) string {
+	str, _ := v.(string)
+	return str
+}
diff --git a/common/ctxkey/key.go b/common/ctxkey/key.go
new file mode 100644
index 00000000..115558a5
--- /dev/null
+++ b/common/ctxkey/key.go
@@ -0,0 +1,24 @@
+package ctxkey
+
+const (
+	Config            = "config"
+	Id                = "id"
+	Username          = "username"
+	Role              = "role"
+	Status            = "status"
+	Channel           = "channel"
+	ChannelId         = "channel_id"
+	SpecificChannelId = "specific_channel_id"
+	RequestModel      = "request_model"
+	ConvertedRequest  = "converted_request"
+	OriginalModel     = "original_model"
+	Group             = "group"
+	ModelMapping      = "model_mapping"
+	ChannelName       = "channel_name"
+	TokenId           = "token_id"
+	TokenName         = "token_name"
+	BaseURL           = "base_url"
+	AvailableModels   = "available_models"
+	KeyRequestBody    = "key_request_body"
+	SystemPrompt      = "system_prompt"
+)
diff --git a/common/gin.go b/common/gin.go
index b6ef96a6..815b4ee5 100644
--- a/common/gin.go
+++ b/common/gin.go
@@ -4,14 +4,13 @@ import (
 	"bytes"
 	"encoding/json"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"io"
 	"strings"
 )
 
-const KeyRequestBody = "key_request_body"
-
 func GetRequestBody(c *gin.Context) ([]byte, error) {
-	requestBody, _ := c.Get(KeyRequestBody)
+	requestBody, _ := c.Get(ctxkey.KeyRequestBody)
 	if requestBody != nil {
 		return requestBody.([]byte), nil
 	}
@@ -20,7 +19,7 @@ func GetRequestBody(c *gin.Context) ([]byte, error) {
 		return nil, err
 	}
 	_ = c.Request.Body.Close()
-	c.Set(KeyRequestBody, requestBody)
+	c.Set(ctxkey.KeyRequestBody, requestBody)
 	return requestBody.([]byte), nil
 }
 
@@ -32,15 +31,15 @@ func UnmarshalBodyReusable(c *gin.Context, v any) error {
 	contentType := c.Request.Header.Get("Content-Type")
 	if strings.HasPrefix(contentType, "application/json") {
 		err = json.Unmarshal(requestBody, &v)
+		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
 	} else {
-		// skip for now
-		// TODO: someday non json request have variant model, we will need to implementation this
+		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
+		err = c.ShouldBind(&v)
 	}
 	if err != nil {
 		return err
 	}
 	// Reset request body
-	c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
 	return nil
 }
 
diff --git a/common/helper/helper.go b/common/helper/helper.go
index db41ac74..df7b0a5f 100644
--- a/common/helper/helper.go
+++ b/common/helper/helper.go
@@ -2,16 +2,15 @@ package helper
 
 import (
 	"fmt"
-	"github.com/google/uuid"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/random"
 	"html/template"
 	"log"
-	"math/rand"
 	"net"
 	"os/exec"
 	"runtime"
 	"strconv"
 	"strings"
-	"time"
 )
 
 func OpenBrowser(url string) {
@@ -79,31 +78,6 @@ func Bytes2Size(num int64) string {
 	return numStr + " " + unit
 }
 
-func Seconds2Time(num int) (time string) {
-	if num/31104000 > 0 {
-		time += strconv.Itoa(num/31104000) + " 年 "
-		num %= 31104000
-	}
-	if num/2592000 > 0 {
-		time += strconv.Itoa(num/2592000) + " 个月 "
-		num %= 2592000
-	}
-	if num/86400 > 0 {
-		time += strconv.Itoa(num/86400) + " 天 "
-		num %= 86400
-	}
-	if num/3600 > 0 {
-		time += strconv.Itoa(num/3600) + " 小时 "
-		num %= 3600
-	}
-	if num/60 > 0 {
-		time += strconv.Itoa(num/60) + " 分钟 "
-		num %= 60
-	}
-	time += strconv.Itoa(num) + " 秒"
-	return
-}
-
 func Interface2String(inter interface{}) string {
 	switch inter := inter.(type) {
 	case string:
@@ -128,65 +102,13 @@ func IntMax(a int, b int) int {
 	}
 }
 
-func GetUUID() string {
-	code := uuid.New().String()
-	code = strings.Replace(code, "-", "", -1)
-	return code
-}
-
-const keyChars = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
-const keyNumbers = "0123456789"
-
-func init() {
-	rand.Seed(time.Now().UnixNano())
-}
-
-func GenerateKey() string {
-	rand.Seed(time.Now().UnixNano())
-	key := make([]byte, 48)
-	for i := 0; i < 16; i++ {
-		key[i] = keyChars[rand.Intn(len(keyChars))]
-	}
-	uuid_ := GetUUID()
-	for i := 0; i < 32; i++ {
-		c := uuid_[i]
-		if i%2 == 0 && c >= 'a' && c <= 'z' {
-			c = c - 'a' + 'A'
-		}
-		key[i+16] = c
-	}
-	return string(key)
-}
-
-func GetRandomString(length int) string {
-	rand.Seed(time.Now().UnixNano())
-	key := make([]byte, length)
-	for i := 0; i < length; i++ {
-		key[i] = keyChars[rand.Intn(len(keyChars))]
-	}
-	return string(key)
-}
-
-func GetRandomNumberString(length int) string {
-	rand.Seed(time.Now().UnixNano())
-	key := make([]byte, length)
-	for i := 0; i < length; i++ {
-		key[i] = keyNumbers[rand.Intn(len(keyNumbers))]
-	}
-	return string(key)
-}
-
-func GetTimestamp() int64 {
-	return time.Now().Unix()
-}
-
-func GetTimeString() string {
-	now := time.Now()
-	return fmt.Sprintf("%s%d", now.Format("20060102150405"), now.UnixNano()%1e9)
-}
-
 func GenRequestID() string {
-	return GetTimeString() + GetRandomNumberString(8)
+	return GetTimeString() + random.GetRandomNumberString(8)
+}
+
+func GetResponseID(c *gin.Context) string {
+	logID := c.GetString(RequestIdKey)
+	return fmt.Sprintf("chatcmpl-%s", logID)
 }
 
 func Max(a int, b int) int {
@@ -215,3 +137,23 @@ func String2Int(str string) int {
 	}
 	return num
 }
+
+func Float64PtrMax(p *float64, maxValue float64) *float64 {
+	if p == nil {
+		return nil
+	}
+	if *p > maxValue {
+		return &maxValue
+	}
+	return p
+}
+
+func Float64PtrMin(p *float64, minValue float64) *float64 {
+	if p == nil {
+		return nil
+	}
+	if *p < minValue {
+		return &minValue
+	}
+	return p
+}
diff --git a/common/helper/key.go b/common/helper/key.go
new file mode 100644
index 00000000..17aee2e0
--- /dev/null
+++ b/common/helper/key.go
@@ -0,0 +1,5 @@
+package helper
+
+const (
+	RequestIdKey = "X-Oneapi-Request-Id"
+)
diff --git a/common/helper/time.go b/common/helper/time.go
new file mode 100644
index 00000000..302746db
--- /dev/null
+++ b/common/helper/time.go
@@ -0,0 +1,15 @@
+package helper
+
+import (
+	"fmt"
+	"time"
+)
+
+func GetTimestamp() int64 {
+	return time.Now().Unix()
+}
+
+func GetTimeString() string {
+	now := time.Now()
+	return fmt.Sprintf("%s%d", now.Format("20060102150405"), now.UnixNano()%1e9)
+}
diff --git a/common/image/image.go b/common/image/image.go
index de8fefd3..beebd0c6 100644
--- a/common/image/image.go
+++ b/common/image/image.go
@@ -3,6 +3,7 @@ package image
 import (
 	"bytes"
 	"encoding/base64"
+	"github.com/songquanpeng/one-api/common/client"
 	"image"
 	_ "image/gif"
 	_ "image/jpeg"
@@ -16,10 +17,10 @@ import (
 )
 
 // Regex to match data URL pattern
-var	dataURLPattern = regexp.MustCompile(`data:image/([^;]+);base64,(.*)`)
+var dataURLPattern = regexp.MustCompile(`data:image/([^;]+);base64,(.*)`)
 
 func IsImageUrl(url string) (bool, error) {
-	resp, err := http.Head(url)
+	resp, err := client.UserContentRequestHTTPClient.Head(url)
 	if err != nil {
 		return false, err
 	}
@@ -34,7 +35,7 @@ func GetImageSizeFromUrl(url string) (width int, height int, err error) {
 	if !isImage {
 		return
 	}
-	resp, err := http.Get(url)
+	resp, err := client.UserContentRequestHTTPClient.Get(url)
 	if err != nil {
 		return
 	}
diff --git a/common/image/image_test.go b/common/image/image_test.go
index 15ed78bc..5b669b51 100644
--- a/common/image/image_test.go
+++ b/common/image/image_test.go
@@ -2,6 +2,7 @@ package image_test
 
 import (
 	"encoding/base64"
+	"github.com/songquanpeng/one-api/common/client"
 	"image"
 	_ "image/gif"
 	_ "image/jpeg"
@@ -44,6 +45,11 @@ var (
 	}
 )
 
+func TestMain(m *testing.M) {
+	client.Init()
+	m.Run()
+}
+
 func TestDecode(t *testing.T) {
 	// Bytes read: varies sometimes
 	// jpeg: 1063892
diff --git a/common/init.go b/common/init.go
index b392bfee..6fd84764 100644
--- a/common/init.go
+++ b/common/init.go
@@ -24,7 +24,7 @@ func printHelp() {
 	fmt.Println("Usage: one-api [--port <port>] [--log-dir <log directory>] [--version] [--help]")
 }
 
-func init() {
+func Init() {
 	flag.Parse()
 
 	if *PrintVersion {
diff --git a/common/logger/constants.go b/common/logger/constants.go
index 78d32062..49df31ec 100644
--- a/common/logger/constants.go
+++ b/common/logger/constants.go
@@ -1,7 +1,3 @@
 package logger
 
-const (
-	RequestIdKey = "X-Oneapi-Request-Id"
-)
-
 var LogDir string
diff --git a/common/logger/logger.go b/common/logger/logger.go
index 957d8a11..d1022932 100644
--- a/common/logger/logger.go
+++ b/common/logger/logger.go
@@ -3,15 +3,16 @@ package logger
 import (
 	"context"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/helper"
 	"io"
 	"log"
 	"os"
 	"path/filepath"
 	"sync"
 	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
 )
 
 const (
@@ -21,28 +22,25 @@ const (
 	loggerError = "ERR"
 )
 
-var setupLogLock sync.Mutex
-var setupLogWorking bool
+var setupLogOnce sync.Once
 
 func SetupLogger() {
-	if LogDir != "" {
-		ok := setupLogLock.TryLock()
-		if !ok {
-			log.Println("setup log is already working")
-			return
+	setupLogOnce.Do(func() {
+		if LogDir != "" {
+			var logPath string
+			if config.OnlyOneLogFile {
+				logPath = filepath.Join(LogDir, "oneapi.log")
+			} else {
+				logPath = filepath.Join(LogDir, fmt.Sprintf("oneapi-%s.log", time.Now().Format("20060102")))
+			}
+			fd, err := os.OpenFile(logPath, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644)
+			if err != nil {
+				log.Fatal("failed to open log file")
+			}
+			gin.DefaultWriter = io.MultiWriter(os.Stdout, fd)
+			gin.DefaultErrorWriter = io.MultiWriter(os.Stderr, fd)
 		}
-		defer func() {
-			setupLogLock.Unlock()
-			setupLogWorking = false
-		}()
-		logPath := filepath.Join(LogDir, fmt.Sprintf("oneapi-%s.log", time.Now().Format("20060102")))
-		fd, err := os.OpenFile(logPath, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644)
-		if err != nil {
-			log.Fatal("failed to open log file")
-		}
-		gin.DefaultWriter = io.MultiWriter(os.Stdout, fd)
-		gin.DefaultErrorWriter = io.MultiWriter(os.Stderr, fd)
-	}
+	})
 }
 
 func SysLog(s string) {
@@ -50,11 +48,19 @@ func SysLog(s string) {
 	_, _ = fmt.Fprintf(gin.DefaultWriter, "[SYS] %v | %s \n", t.Format("2006/01/02 - 15:04:05"), s)
 }
 
+func SysLogf(format string, a ...any) {
+	SysLog(fmt.Sprintf(format, a...))
+}
+
 func SysError(s string) {
 	t := time.Now()
 	_, _ = fmt.Fprintf(gin.DefaultErrorWriter, "[SYS] %v | %s \n", t.Format("2006/01/02 - 15:04:05"), s)
 }
 
+func SysErrorf(format string, a ...any) {
+	SysError(fmt.Sprintf(format, a...))
+}
+
 func Debug(ctx context.Context, msg string) {
 	if config.DebugEnabled {
 		logHelper(ctx, loggerDEBUG, msg)
@@ -94,18 +100,13 @@ func logHelper(ctx context.Context, level string, msg string) {
 	if level == loggerINFO {
 		writer = gin.DefaultWriter
 	}
-	id := ctx.Value(RequestIdKey)
+	id := ctx.Value(helper.RequestIdKey)
 	if id == nil {
 		id = helper.GenRequestID()
 	}
 	now := time.Now()
 	_, _ = fmt.Fprintf(writer, "[%s] %v | %s | %s \n", level, now.Format("2006/01/02 - 15:04:05"), id, msg)
-	if !setupLogWorking {
-		setupLogWorking = true
-		go func() {
-			SetupLogger()
-		}()
-	}
+	SetupLogger()
 }
 
 func FatalLog(v ...any) {
diff --git a/common/message/email.go b/common/message/email.go
index b06782db..187ac8c3 100644
--- a/common/message/email.go
+++ b/common/message/email.go
@@ -6,11 +6,16 @@ import (
 	"encoding/base64"
 	"fmt"
 	"github.com/songquanpeng/one-api/common/config"
+	"net"
 	"net/smtp"
 	"strings"
 	"time"
 )
 
+func shouldAuth() bool {
+	return config.SMTPAccount != "" || config.SMTPToken != ""
+}
+
 func SendEmail(subject string, receiver string, content string) error {
 	if receiver == "" {
 		return fmt.Errorf("receiver is empty")
@@ -41,16 +46,24 @@ func SendEmail(subject string, receiver string, content string) error {
 		"Date: %s\r\n"+
 		"Content-Type: text/html; charset=UTF-8\r\n\r\n%s\r\n",
 		receiver, config.SystemName, config.SMTPFrom, encodedSubject, messageId, time.Now().Format(time.RFC1123Z), content))
+
 	auth := smtp.PlainAuth("", config.SMTPAccount, config.SMTPToken, config.SMTPServer)
 	addr := fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort)
 	to := strings.Split(receiver, ";")
 
-	if config.SMTPPort == 465 {
-		tlsConfig := &tls.Config{
-			InsecureSkipVerify: true,
-			ServerName:         config.SMTPServer,
+	if config.SMTPPort == 465 || !shouldAuth() {
+		// need advanced client
+		var conn net.Conn
+		var err error
+		if config.SMTPPort == 465 {
+			tlsConfig := &tls.Config{
+				InsecureSkipVerify: true,
+				ServerName:         config.SMTPServer,
+			}
+			conn, err = tls.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort), tlsConfig)
+		} else {
+			conn, err = net.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort))
 		}
-		conn, err := tls.Dial("tcp", fmt.Sprintf("%s:%d", config.SMTPServer, config.SMTPPort), tlsConfig)
 		if err != nil {
 			return err
 		}
@@ -59,8 +72,10 @@ func SendEmail(subject string, receiver string, content string) error {
 			return err
 		}
 		defer client.Close()
-		if err = client.Auth(auth); err != nil {
-			return err
+		if shouldAuth() {
+			if err = client.Auth(auth); err != nil {
+				return err
+			}
 		}
 		if err = client.Mail(config.SMTPFrom); err != nil {
 			return err
diff --git a/common/model-ratio.go b/common/model-ratio.go
deleted file mode 100644
index 2c608302..00000000
--- a/common/model-ratio.go
+++ /dev/null
@@ -1,257 +0,0 @@
-package common
-
-import (
-	"encoding/json"
-	"github.com/songquanpeng/one-api/common/logger"
-	"strings"
-)
-
-const (
-	USD2RMB = 7
-	USD     = 500 // $0.002 = 1 -> $1 = 500
-	RMB     = USD / USD2RMB
-)
-
-// ModelRatio
-// https://platform.openai.com/docs/models/model-endpoint-compatibility
-// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
-// https://openai.com/pricing
-// 1 === $0.002 / 1K tokens
-// 1 === ￥0.014 / 1k tokens
-var ModelRatio = map[string]float64{
-	// https://openai.com/pricing
-	"gpt-4":                   15,
-	"gpt-4-0314":              15,
-	"gpt-4-0613":              15,
-	"gpt-4-32k":               30,
-	"gpt-4-32k-0314":          30,
-	"gpt-4-32k-0613":          30,
-	"gpt-4-1106-preview":      5,    // $0.01 / 1K tokens
-	"gpt-4-0125-preview":      5,    // $0.01 / 1K tokens
-	"gpt-4-turbo-preview":     5,    // $0.01 / 1K tokens
-	"gpt-4-vision-preview":    5,    // $0.01 / 1K tokens
-	"gpt-3.5-turbo":           0.25, // $0.0005 / 1K tokens
-	"gpt-3.5-turbo-0301":      0.75,
-	"gpt-3.5-turbo-0613":      0.75,
-	"gpt-3.5-turbo-16k":       1.5, // $0.003 / 1K tokens
-	"gpt-3.5-turbo-16k-0613":  1.5,
-	"gpt-3.5-turbo-instruct":  0.75, // $0.0015 / 1K tokens
-	"gpt-3.5-turbo-1106":      0.5,  // $0.001 / 1K tokens
-	"gpt-3.5-turbo-0125":      0.25, // $0.0005 / 1K tokens
-	"davinci-002":             1,    // $0.002 / 1K tokens
-	"babbage-002":             0.2,  // $0.0004 / 1K tokens
-	"text-ada-001":            0.2,
-	"text-babbage-001":        0.25,
-	"text-curie-001":          1,
-	"text-davinci-002":        10,
-	"text-davinci-003":        10,
-	"text-davinci-edit-001":   10,
-	"code-davinci-edit-001":   10,
-	"whisper-1":               15,  // $0.006 / minute -> $0.006 / 150 words -> $0.006 / 200 tokens -> $0.03 / 1k tokens
-	"tts-1":                   7.5, // $0.015 / 1K characters
-	"tts-1-1106":              7.5,
-	"tts-1-hd":                15, // $0.030 / 1K characters
-	"tts-1-hd-1106":           15,
-	"davinci":                 10,
-	"curie":                   10,
-	"babbage":                 10,
-	"ada":                     10,
-	"text-embedding-ada-002":  0.05,
-	"text-embedding-3-small":  0.01,
-	"text-embedding-3-large":  0.065,
-	"text-search-ada-doc-001": 10,
-	"text-moderation-stable":  0.1,
-	"text-moderation-latest":  0.1,
-	"dall-e-2":                8,  // $0.016 - $0.020 / image
-	"dall-e-3":                20, // $0.040 - $0.120 / image
-	// https://www.anthropic.com/api#pricing
-	"claude-instant-1.2":       0.8 / 1000 * USD,
-	"claude-2.0":               8.0 / 1000 * USD,
-	"claude-2.1":               8.0 / 1000 * USD,
-	"claude-3-haiku-20240307":  0.25 / 1000 * USD,
-	"claude-3-sonnet-20240229": 3.0 / 1000 * USD,
-	"claude-3-opus-20240229":   15.0 / 1000 * USD,
-	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
-	"ERNIE-Bot":       0.8572,     // ￥0.012 / 1k tokens
-	"ERNIE-Bot-turbo": 0.5715,     // ￥0.008 / 1k tokens
-	"ERNIE-Bot-4":     0.12 * RMB, // ￥0.12 / 1k tokens
-	"ERNIE-Bot-8k":    0.024 * RMB,
-	"Embedding-V1":    0.1429, // ￥0.002 / 1k tokens
-	"bge-large-zh":    0.002 * RMB,
-	"bge-large-en":    0.002 * RMB,
-	"bge-large-8k":    0.002 * RMB,
-	// https://ai.google.dev/pricing
-	"PaLM-2":            1,
-	"gemini-pro":        1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
-	"gemini-pro-vision": 1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
-	// https://open.bigmodel.cn/pricing
-	"glm-4":                     0.1 * RMB,
-	"glm-4v":                    0.1 * RMB,
-	"glm-3-turbo":               0.005 * RMB,
-	"chatglm_turbo":             0.3572, // ￥0.005 / 1k tokens
-	"chatglm_pro":               0.7143, // ￥0.01 / 1k tokens
-	"chatglm_std":               0.3572, // ￥0.005 / 1k tokens
-	"chatglm_lite":              0.1429, // ￥0.002 / 1k tokens
-	"qwen-turbo":                0.5715, // ￥0.008 / 1k tokens  // https://help.aliyun.com/zh/dashscope/developer-reference/tongyi-thousand-questions-metering-and-billing
-	"qwen-plus":                 1.4286, // ￥0.02 / 1k tokens
-	"qwen-max":                  1.4286, // ￥0.02 / 1k tokens
-	"qwen-max-longcontext":      1.4286, // ￥0.02 / 1k tokens
-	"text-embedding-v1":         0.05,   // ￥0.0007 / 1k tokens
-	"SparkDesk":                 1.2858, // ￥0.018 / 1k tokens
-	"SparkDesk-v1.1":            1.2858, // ￥0.018 / 1k tokens
-	"SparkDesk-v2.1":            1.2858, // ￥0.018 / 1k tokens
-	"SparkDesk-v3.1":            1.2858, // ￥0.018 / 1k tokens
-	"SparkDesk-v3.5":            1.2858, // ￥0.018 / 1k tokens
-	"360GPT_S2_V9":              0.8572, // ¥0.012 / 1k tokens
-	"embedding-bert-512-v1":     0.0715, // ¥0.001 / 1k tokens
-	"embedding_s1_v1":           0.0715, // ¥0.001 / 1k tokens
-	"semantic_similarity_s1_v1": 0.0715, // ¥0.001 / 1k tokens
-	"hunyuan":                   7.143,  // ¥0.1 / 1k tokens  // https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
-	"ChatStd":                   0.01 * RMB,
-	"ChatPro":                   0.1 * RMB,
-	// https://platform.moonshot.cn/pricing
-	"moonshot-v1-8k":   0.012 * RMB,
-	"moonshot-v1-32k":  0.024 * RMB,
-	"moonshot-v1-128k": 0.06 * RMB,
-	// https://platform.baichuan-ai.com/price
-	"Baichuan2-Turbo":      0.008 * RMB,
-	"Baichuan2-Turbo-192k": 0.016 * RMB,
-	"Baichuan2-53B":        0.02 * RMB,
-	// https://api.minimax.chat/document/price
-	"abab6-chat":    0.1 * RMB,
-	"abab5.5-chat":  0.015 * RMB,
-	"abab5.5s-chat": 0.005 * RMB,
-	// https://docs.mistral.ai/platform/pricing/
-	"open-mistral-7b":       0.25 / 1000 * USD,
-	"open-mixtral-8x7b":     0.7 / 1000 * USD,
-	"mistral-small-latest":  2.0 / 1000 * USD,
-	"mistral-medium-latest": 2.7 / 1000 * USD,
-	"mistral-large-latest":  8.0 / 1000 * USD,
-	"mistral-embed":         0.1 / 1000 * USD,
-	// https://wow.groq.com/
-	"llama2-70b-4096":    0.7 / 1000 * USD,
-	"llama2-7b-2048":     0.1 / 1000 * USD,
-	"mixtral-8x7b-32768": 0.27 / 1000 * USD,
-	"gemma-7b-it":        0.1 / 1000 * USD,
-	// https://platform.lingyiwanwu.com/docs#-计费单元
-	"yi-34b-chat-0205": 2.5 / 1000 * RMB,
-	"yi-34b-chat-200k": 12.0 / 1000 * RMB,
-	"yi-vl-plus":       6.0 / 1000 * RMB,
-}
-
-var CompletionRatio = map[string]float64{}
-
-var DefaultModelRatio map[string]float64
-var DefaultCompletionRatio map[string]float64
-
-func init() {
-	DefaultModelRatio = make(map[string]float64)
-	for k, v := range ModelRatio {
-		DefaultModelRatio[k] = v
-	}
-	DefaultCompletionRatio = make(map[string]float64)
-	for k, v := range CompletionRatio {
-		DefaultCompletionRatio[k] = v
-	}
-}
-
-func AddNewMissingRatio(oldRatio string) string {
-	newRatio := make(map[string]float64)
-	err := json.Unmarshal([]byte(oldRatio), &newRatio)
-	if err != nil {
-		logger.SysError("error unmarshalling old ratio: " + err.Error())
-		return oldRatio
-	}
-	for k, v := range DefaultModelRatio {
-		if _, ok := newRatio[k]; !ok {
-			newRatio[k] = v
-		}
-	}
-	jsonBytes, err := json.Marshal(newRatio)
-	if err != nil {
-		logger.SysError("error marshalling new ratio: " + err.Error())
-		return oldRatio
-	}
-	return string(jsonBytes)
-}
-
-func ModelRatio2JSONString() string {
-	jsonBytes, err := json.Marshal(ModelRatio)
-	if err != nil {
-		logger.SysError("error marshalling model ratio: " + err.Error())
-	}
-	return string(jsonBytes)
-}
-
-func UpdateModelRatioByJSONString(jsonStr string) error {
-	ModelRatio = make(map[string]float64)
-	return json.Unmarshal([]byte(jsonStr), &ModelRatio)
-}
-
-func GetModelRatio(name string) float64 {
-	if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
-		name = strings.TrimSuffix(name, "-internet")
-	}
-	ratio, ok := ModelRatio[name]
-	if !ok {
-		ratio, ok = DefaultModelRatio[name]
-	}
-	if !ok {
-		logger.SysError("model ratio not found: " + name)
-		return 30
-	}
-	return ratio
-}
-
-func CompletionRatio2JSONString() string {
-	jsonBytes, err := json.Marshal(CompletionRatio)
-	if err != nil {
-		logger.SysError("error marshalling completion ratio: " + err.Error())
-	}
-	return string(jsonBytes)
-}
-
-func UpdateCompletionRatioByJSONString(jsonStr string) error {
-	CompletionRatio = make(map[string]float64)
-	return json.Unmarshal([]byte(jsonStr), &CompletionRatio)
-}
-
-func GetCompletionRatio(name string) float64 {
-	if ratio, ok := CompletionRatio[name]; ok {
-		return ratio
-	}
-	if ratio, ok := DefaultCompletionRatio[name]; ok {
-		return ratio
-	}
-	if strings.HasPrefix(name, "gpt-3.5") {
-		if name == "gpt-3.5-turbo" || strings.HasSuffix(name, "0125") {
-			// https://openai.com/blog/new-embedding-models-and-api-updates
-			// Updated GPT-3.5 Turbo model and lower pricing
-			return 3
-		}
-		if strings.HasSuffix(name, "1106") {
-			return 2
-		}
-		return 4.0 / 3.0
-	}
-	if strings.HasPrefix(name, "gpt-4") {
-		if strings.HasSuffix(name, "preview") {
-			return 3
-		}
-		return 2
-	}
-	if strings.HasPrefix(name, "claude-3") {
-		return 5
-	}
-	if strings.HasPrefix(name, "claude-") {
-		return 3
-	}
-	if strings.HasPrefix(name, "mistral-") {
-		return 3
-	}
-	switch name {
-	case "llama2-70b-4096":
-		return 0.8 / 0.7
-	}
-	return 1
-}
diff --git a/common/network/ip.go b/common/network/ip.go
new file mode 100644
index 00000000..0fbe5e6f
--- /dev/null
+++ b/common/network/ip.go
@@ -0,0 +1,52 @@
+package network
+
+import (
+	"context"
+	"fmt"
+	"github.com/songquanpeng/one-api/common/logger"
+	"net"
+	"strings"
+)
+
+func splitSubnets(subnets string) []string {
+	res := strings.Split(subnets, ",")
+	for i := 0; i < len(res); i++ {
+		res[i] = strings.TrimSpace(res[i])
+	}
+	return res
+}
+
+func isValidSubnet(subnet string) error {
+	_, _, err := net.ParseCIDR(subnet)
+	if err != nil {
+		return fmt.Errorf("failed to parse subnet: %w", err)
+	}
+	return nil
+}
+
+func isIpInSubnet(ctx context.Context, ip string, subnet string) bool {
+	_, ipNet, err := net.ParseCIDR(subnet)
+	if err != nil {
+		logger.Errorf(ctx, "failed to parse subnet: %s", err.Error())
+		return false
+	}
+	return ipNet.Contains(net.ParseIP(ip))
+}
+
+func IsValidSubnets(subnets string) error {
+	for _, subnet := range splitSubnets(subnets) {
+		if err := isValidSubnet(subnet); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func IsIpInSubnets(ctx context.Context, ip string, subnets string) bool {
+	for _, subnet := range splitSubnets(subnets) {
+		if isIpInSubnet(ctx, ip, subnet) {
+			return true
+		}
+	}
+	return false
+}
diff --git a/common/network/ip_test.go b/common/network/ip_test.go
new file mode 100644
index 00000000..6c593458
--- /dev/null
+++ b/common/network/ip_test.go
@@ -0,0 +1,19 @@
+package network
+
+import (
+	"context"
+	"testing"
+
+	. "github.com/smartystreets/goconvey/convey"
+)
+
+func TestIsIpInSubnet(t *testing.T) {
+	ctx := context.Background()
+	ip1 := "192.168.0.5"
+	ip2 := "125.216.250.89"
+	subnet := "192.168.0.0/24"
+	Convey("TestIsIpInSubnet", t, func() {
+		So(isIpInSubnet(ctx, ip1, subnet), ShouldBeTrue)
+		So(isIpInSubnet(ctx, ip2, subnet), ShouldBeFalse)
+	})
+}
diff --git a/common/random.go b/common/random.go
deleted file mode 100644
index 44bd2856..00000000
--- a/common/random.go
+++ /dev/null
@@ -1,8 +0,0 @@
-package common
-
-import "math/rand"
-
-// RandRange returns a random number between min and max (max is not included)
-func RandRange(min, max int) int {
-	return min + rand.Intn(max-min)
-}
diff --git a/common/random/main.go b/common/random/main.go
new file mode 100644
index 00000000..dbb772cd
--- /dev/null
+++ b/common/random/main.go
@@ -0,0 +1,61 @@
+package random
+
+import (
+	"github.com/google/uuid"
+	"math/rand"
+	"strings"
+	"time"
+)
+
+func GetUUID() string {
+	code := uuid.New().String()
+	code = strings.Replace(code, "-", "", -1)
+	return code
+}
+
+const keyChars = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"
+const keyNumbers = "0123456789"
+
+func init() {
+	rand.Seed(time.Now().UnixNano())
+}
+
+func GenerateKey() string {
+	rand.Seed(time.Now().UnixNano())
+	key := make([]byte, 48)
+	for i := 0; i < 16; i++ {
+		key[i] = keyChars[rand.Intn(len(keyChars))]
+	}
+	uuid_ := GetUUID()
+	for i := 0; i < 32; i++ {
+		c := uuid_[i]
+		if i%2 == 0 && c >= 'a' && c <= 'z' {
+			c = c - 'a' + 'A'
+		}
+		key[i+16] = c
+	}
+	return string(key)
+}
+
+func GetRandomString(length int) string {
+	rand.Seed(time.Now().UnixNano())
+	key := make([]byte, length)
+	for i := 0; i < length; i++ {
+		key[i] = keyChars[rand.Intn(len(keyChars))]
+	}
+	return string(key)
+}
+
+func GetRandomNumberString(length int) string {
+	rand.Seed(time.Now().UnixNano())
+	key := make([]byte, length)
+	for i := 0; i < length; i++ {
+		key[i] = keyNumbers[rand.Intn(len(keyNumbers))]
+	}
+	return string(key)
+}
+
+// RandRange returns a random number between min and max (max is not included)
+func RandRange(min, max int) int {
+	return min + rand.Intn(max-min)
+}
diff --git a/common/redis.go b/common/redis.go
index f3205567..55d4931c 100644
--- a/common/redis.go
+++ b/common/redis.go
@@ -2,13 +2,15 @@ package common
 
 import (
 	"context"
+	"os"
+	"strings"
+	"time"
+
 	"github.com/go-redis/redis/v8"
 	"github.com/songquanpeng/one-api/common/logger"
-	"os"
-	"time"
 )
 
-var RDB *redis.Client
+var RDB redis.Cmdable
 var RedisEnabled = true
 
 // InitRedisClient This function is called after init()
@@ -23,13 +25,23 @@ func InitRedisClient() (err error) {
 		logger.SysLog("SYNC_FREQUENCY not set, Redis is disabled")
 		return nil
 	}
-	logger.SysLog("Redis is enabled")
-	opt, err := redis.ParseURL(os.Getenv("REDIS_CONN_STRING"))
-	if err != nil {
-		logger.FatalLog("failed to parse Redis connection string: " + err.Error())
+	redisConnString := os.Getenv("REDIS_CONN_STRING")
+	if os.Getenv("REDIS_MASTER_NAME") == "" {
+		logger.SysLog("Redis is enabled")
+		opt, err := redis.ParseURL(redisConnString)
+		if err != nil {
+			logger.FatalLog("failed to parse Redis connection string: " + err.Error())
+		}
+		RDB = redis.NewClient(opt)
+	} else {
+		// cluster mode
+		logger.SysLog("Redis cluster mode enabled")
+		RDB = redis.NewUniversalClient(&redis.UniversalOptions{
+			Addrs:      strings.Split(redisConnString, ","),
+			Password:   os.Getenv("REDIS_PASSWORD"),
+			MasterName: os.Getenv("REDIS_MASTER_NAME"),
+		})
 	}
-	RDB = redis.NewClient(opt)
-
 	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
 	defer cancel()
 
diff --git a/common/render/render.go b/common/render/render.go
new file mode 100644
index 00000000..e565c0b7
--- /dev/null
+++ b/common/render/render.go
@@ -0,0 +1,30 @@
+package render
+
+import (
+	"encoding/json"
+	"fmt"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+)
+
+func StringData(c *gin.Context, str string) {
+	str = strings.TrimPrefix(str, "data: ")
+	str = strings.TrimSuffix(str, "\r")
+	c.Render(-1, common.CustomEvent{Data: "data: " + str})
+	c.Writer.Flush()
+}
+
+func ObjectData(c *gin.Context, object interface{}) error {
+	jsonData, err := json.Marshal(object)
+	if err != nil {
+		return fmt.Errorf("error marshalling object: %w", err)
+	}
+	StringData(c, string(jsonData))
+	return nil
+}
+
+func Done(c *gin.Context) {
+	StringData(c, "[DONE]")
+}
diff --git a/controller/github.go b/controller/auth/github.go
similarity index 94%
rename from controller/github.go
rename to controller/auth/github.go
index 7d7fa106..15542655 100644
--- a/controller/github.go
+++ b/controller/auth/github.go
@@ -1,4 +1,4 @@
-package controller
+package auth
 
 import (
 	"bytes"
@@ -7,10 +7,10 @@ import (
 	"fmt"
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/controller"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
@@ -133,8 +133,8 @@ func GitHubOAuth(c *gin.Context) {
 				user.DisplayName = "GitHub User"
 			}
 			user.Email = githubUser.Email
-			user.Role = common.RoleCommonUser
-			user.Status = common.UserStatusEnabled
+			user.Role = model.RoleCommonUser
+			user.Status = model.UserStatusEnabled
 
 			if err := user.Insert(0); err != nil {
 				c.JSON(http.StatusOK, gin.H{
@@ -152,14 +152,14 @@ func GitHubOAuth(c *gin.Context) {
 		}
 	}
 
-	if user.Status != common.UserStatusEnabled {
+	if user.Status != model.UserStatusEnabled {
 		c.JSON(http.StatusOK, gin.H{
 			"message": "用户已被封禁",
 			"success": false,
 		})
 		return
 	}
-	setupLogin(&user, c)
+	controller.SetupLogin(&user, c)
 }
 
 func GitHubBind(c *gin.Context) {
@@ -219,7 +219,7 @@ func GitHubBind(c *gin.Context) {
 
 func GenerateOAuthCode(c *gin.Context) {
 	session := sessions.Default(c)
-	state := helper.GetRandomString(12)
+	state := random.GetRandomString(12)
 	session.Set("oauth_state", state)
 	err := session.Save()
 	if err != nil {
diff --git a/controller/auth/lark.go b/controller/auth/lark.go
new file mode 100644
index 00000000..39088b3c
--- /dev/null
+++ b/controller/auth/lark.go
@@ -0,0 +1,200 @@
+package auth
+
+import (
+	"bytes"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"github.com/gin-contrib/sessions"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/controller"
+	"github.com/songquanpeng/one-api/model"
+	"net/http"
+	"strconv"
+	"time"
+)
+
+type LarkOAuthResponse struct {
+	AccessToken string `json:"access_token"`
+}
+
+type LarkUser struct {
+	Name   string `json:"name"`
+	OpenID string `json:"open_id"`
+}
+
+func getLarkUserInfoByCode(code string) (*LarkUser, error) {
+	if code == "" {
+		return nil, errors.New("无效的参数")
+	}
+	values := map[string]string{
+		"client_id":     config.LarkClientId,
+		"client_secret": config.LarkClientSecret,
+		"code":          code,
+		"grant_type":    "authorization_code",
+		"redirect_uri":  fmt.Sprintf("%s/oauth/lark", config.ServerAddress),
+	}
+	jsonData, err := json.Marshal(values)
+	if err != nil {
+		return nil, err
+	}
+	req, err := http.NewRequest("POST", "https://open.feishu.cn/open-apis/authen/v2/oauth/token", bytes.NewBuffer(jsonData))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("Accept", "application/json")
+	client := http.Client{
+		Timeout: 5 * time.Second,
+	}
+	res, err := client.Do(req)
+	if err != nil {
+		logger.SysLog(err.Error())
+		return nil, errors.New("无法连接至飞书服务器，请稍后重试！")
+	}
+	defer res.Body.Close()
+	var oAuthResponse LarkOAuthResponse
+	err = json.NewDecoder(res.Body).Decode(&oAuthResponse)
+	if err != nil {
+		return nil, err
+	}
+	req, err = http.NewRequest("GET", "https://passport.feishu.cn/suite/passport/oauth/userinfo", nil)
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", oAuthResponse.AccessToken))
+	res2, err := client.Do(req)
+	if err != nil {
+		logger.SysLog(err.Error())
+		return nil, errors.New("无法连接至飞书服务器，请稍后重试！")
+	}
+	var larkUser LarkUser
+	err = json.NewDecoder(res2.Body).Decode(&larkUser)
+	if err != nil {
+		return nil, err
+	}
+	return &larkUser, nil
+}
+
+func LarkOAuth(c *gin.Context) {
+	session := sessions.Default(c)
+	state := c.Query("state")
+	if state == "" || session.Get("oauth_state") == nil || state != session.Get("oauth_state").(string) {
+		c.JSON(http.StatusForbidden, gin.H{
+			"success": false,
+			"message": "state is empty or not same",
+		})
+		return
+	}
+	username := session.Get("username")
+	if username != nil {
+		LarkBind(c)
+		return
+	}
+	code := c.Query("code")
+	larkUser, err := getLarkUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		LarkId: larkUser.OpenID,
+	}
+	if model.IsLarkIdAlreadyTaken(user.LarkId) {
+		err := user.FillUserByLarkId()
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	} else {
+		if config.RegisterEnabled {
+			user.Username = "lark_" + strconv.Itoa(model.GetMaxUserId()+1)
+			if larkUser.Name != "" {
+				user.DisplayName = larkUser.Name
+			} else {
+				user.DisplayName = "Lark User"
+			}
+			user.Role = model.RoleCommonUser
+			user.Status = model.UserStatusEnabled
+
+			if err := user.Insert(0); err != nil {
+				c.JSON(http.StatusOK, gin.H{
+					"success": false,
+					"message": err.Error(),
+				})
+				return
+			}
+		} else {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "管理员关闭了新用户注册",
+			})
+			return
+		}
+	}
+
+	if user.Status != model.UserStatusEnabled {
+		c.JSON(http.StatusOK, gin.H{
+			"message": "用户已被封禁",
+			"success": false,
+		})
+		return
+	}
+	controller.SetupLogin(&user, c)
+}
+
+func LarkBind(c *gin.Context) {
+	code := c.Query("code")
+	larkUser, err := getLarkUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		LarkId: larkUser.OpenID,
+	}
+	if model.IsLarkIdAlreadyTaken(user.LarkId) {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "该飞书账户已被绑定",
+		})
+		return
+	}
+	session := sessions.Default(c)
+	id := session.Get("id")
+	// id := c.GetInt("id")  // critical bug!
+	user.Id = id.(int)
+	err = user.FillUserById()
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user.LarkId = larkUser.OpenID
+	err = user.Update(false)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "bind",
+	})
+	return
+}
diff --git a/controller/auth/oidc.go b/controller/auth/oidc.go
new file mode 100644
index 00000000..7b4ad4b9
--- /dev/null
+++ b/controller/auth/oidc.go
@@ -0,0 +1,225 @@
+package auth
+
+import (
+	"bytes"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"github.com/gin-contrib/sessions"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/controller"
+	"github.com/songquanpeng/one-api/model"
+	"net/http"
+	"strconv"
+	"time"
+)
+
+type OidcResponse struct {
+	AccessToken  string `json:"access_token"`
+	IDToken      string `json:"id_token"`
+	RefreshToken string `json:"refresh_token"`
+	TokenType    string `json:"token_type"`
+	ExpiresIn    int    `json:"expires_in"`
+	Scope        string `json:"scope"`
+}
+
+type OidcUser struct {
+	OpenID            string `json:"sub"`
+	Email             string `json:"email"`
+	Name              string `json:"name"`
+	PreferredUsername string `json:"preferred_username"`
+	Picture           string `json:"picture"`
+}
+
+func getOidcUserInfoByCode(code string) (*OidcUser, error) {
+	if code == "" {
+		return nil, errors.New("无效的参数")
+	}
+	values := map[string]string{
+		"client_id":     config.OidcClientId,
+		"client_secret": config.OidcClientSecret,
+		"code":          code,
+		"grant_type":    "authorization_code",
+		"redirect_uri":  fmt.Sprintf("%s/oauth/oidc", config.ServerAddress),
+	}
+	jsonData, err := json.Marshal(values)
+	if err != nil {
+		return nil, err
+	}
+	req, err := http.NewRequest("POST", config.OidcTokenEndpoint, bytes.NewBuffer(jsonData))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("Accept", "application/json")
+	client := http.Client{
+		Timeout: 5 * time.Second,
+	}
+	res, err := client.Do(req)
+	if err != nil {
+		logger.SysLog(err.Error())
+		return nil, errors.New("无法连接至 OIDC 服务器，请稍后重试！")
+	}
+	defer res.Body.Close()
+	var oidcResponse OidcResponse
+	err = json.NewDecoder(res.Body).Decode(&oidcResponse)
+	if err != nil {
+		return nil, err
+	}
+	req, err = http.NewRequest("GET", config.OidcUserinfoEndpoint, nil)
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Authorization", "Bearer "+oidcResponse.AccessToken)
+	res2, err := client.Do(req)
+	if err != nil {
+		logger.SysLog(err.Error())
+		return nil, errors.New("无法连接至 OIDC 服务器，请稍后重试！")
+	}
+	var oidcUser OidcUser
+	err = json.NewDecoder(res2.Body).Decode(&oidcUser)
+	if err != nil {
+		return nil, err
+	}
+	return &oidcUser, nil
+}
+
+func OidcAuth(c *gin.Context) {
+	session := sessions.Default(c)
+	state := c.Query("state")
+	if state == "" || session.Get("oauth_state") == nil || state != session.Get("oauth_state").(string) {
+		c.JSON(http.StatusForbidden, gin.H{
+			"success": false,
+			"message": "state is empty or not same",
+		})
+		return
+	}
+	username := session.Get("username")
+	if username != nil {
+		OidcBind(c)
+		return
+	}
+	if !config.OidcEnabled {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "管理员未开启通过 OIDC 登录以及注册",
+		})
+		return
+	}
+	code := c.Query("code")
+	oidcUser, err := getOidcUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		OidcId: oidcUser.OpenID,
+	}
+	if model.IsOidcIdAlreadyTaken(user.OidcId) {
+		err := user.FillUserByOidcId()
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	} else {
+		if config.RegisterEnabled {
+			user.Email = oidcUser.Email
+			if oidcUser.PreferredUsername != "" {
+				user.Username = oidcUser.PreferredUsername
+			} else {
+				user.Username = "oidc_" + strconv.Itoa(model.GetMaxUserId()+1)
+			}
+			if oidcUser.Name != "" {
+				user.DisplayName = oidcUser.Name
+			} else {
+				user.DisplayName = "OIDC User"
+			}
+			err := user.Insert(0)
+			if err != nil {
+				c.JSON(http.StatusOK, gin.H{
+					"success": false,
+					"message": err.Error(),
+				})
+				return
+			}
+		} else {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "管理员关闭了新用户注册",
+			})
+			return
+		}
+	}
+
+	if user.Status != model.UserStatusEnabled {
+		c.JSON(http.StatusOK, gin.H{
+			"message": "用户已被封禁",
+			"success": false,
+		})
+		return
+	}
+	controller.SetupLogin(&user, c)
+}
+
+func OidcBind(c *gin.Context) {
+	if !config.OidcEnabled {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "管理员未开启通过 OIDC 登录以及注册",
+		})
+		return
+	}
+	code := c.Query("code")
+	oidcUser, err := getOidcUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		OidcId: oidcUser.OpenID,
+	}
+	if model.IsOidcIdAlreadyTaken(user.OidcId) {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "该 OIDC 账户已被绑定",
+		})
+		return
+	}
+	session := sessions.Default(c)
+	id := session.Get("id")
+	// id := c.GetInt("id")  // critical bug!
+	user.Id = id.(int)
+	err = user.FillUserById()
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user.OidcId = oidcUser.OpenID
+	err = user.Update(false)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "bind",
+	})
+	return
+}
diff --git a/controller/wechat.go b/controller/auth/wechat.go
similarity index 91%
rename from controller/wechat.go
rename to controller/auth/wechat.go
index 74be5604..a561aec0 100644
--- a/controller/wechat.go
+++ b/controller/auth/wechat.go
@@ -1,12 +1,13 @@
-package controller
+package auth
 
 import (
 	"encoding/json"
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/controller"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
@@ -83,8 +84,8 @@ func WeChatAuth(c *gin.Context) {
 		if config.RegisterEnabled {
 			user.Username = "wechat_" + strconv.Itoa(model.GetMaxUserId()+1)
 			user.DisplayName = "WeChat User"
-			user.Role = common.RoleCommonUser
-			user.Status = common.UserStatusEnabled
+			user.Role = model.RoleCommonUser
+			user.Status = model.UserStatusEnabled
 
 			if err := user.Insert(0); err != nil {
 				c.JSON(http.StatusOK, gin.H{
@@ -102,14 +103,14 @@ func WeChatAuth(c *gin.Context) {
 		}
 	}
 
-	if user.Status != common.UserStatusEnabled {
+	if user.Status != model.UserStatusEnabled {
 		c.JSON(http.StatusOK, gin.H{
 			"message": "用户已被封禁",
 			"success": false,
 		})
 		return
 	}
-	setupLogin(&user, c)
+	controller.SetupLogin(&user, c)
 }
 
 func WeChatBind(c *gin.Context) {
@@ -136,7 +137,7 @@ func WeChatBind(c *gin.Context) {
 		})
 		return
 	}
-	id := c.GetInt("id")
+	id := c.GetInt(ctxkey.Id)
 	user := model.User{
 		Id: id,
 	}
diff --git a/controller/billing.go b/controller/billing.go
index dd518678..e837157f 100644
--- a/controller/billing.go
+++ b/controller/billing.go
@@ -3,6 +3,7 @@ package controller
 import (
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/model"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
 )
@@ -14,13 +15,15 @@ func GetSubscription(c *gin.Context) {
 	var token *model.Token
 	var expiredTime int64
 	if config.DisplayTokenStatEnabled {
-		tokenId := c.GetInt("token_id")
+		tokenId := c.GetInt(ctxkey.TokenId)
 		token, err = model.GetTokenById(tokenId)
-		expiredTime = token.ExpiredTime
-		remainQuota = token.RemainQuota
-		usedQuota = token.UsedQuota
+		if err == nil {
+			expiredTime = token.ExpiredTime
+			remainQuota = token.RemainQuota
+			usedQuota = token.UsedQuota
+		}
 	} else {
-		userId := c.GetInt("id")
+		userId := c.GetInt(ctxkey.Id)
 		remainQuota, err = model.GetUserQuota(userId)
 		if err != nil {
 			usedQuota, err = model.GetUserUsedQuota(userId)
@@ -64,11 +67,11 @@ func GetUsage(c *gin.Context) {
 	var err error
 	var token *model.Token
 	if config.DisplayTokenStatEnabled {
-		tokenId := c.GetInt("token_id")
+		tokenId := c.GetInt(ctxkey.TokenId)
 		token, err = model.GetTokenById(tokenId)
 		quota = token.UsedQuota
 	} else {
-		userId := c.GetInt("id")
+		userId := c.GetInt(ctxkey.Id)
 		quota, err = model.GetUserUsedQuota(userId)
 	}
 	if err != nil {
diff --git a/controller/channel-billing.go b/controller/channel-billing.go
index 03c97349..e69cd9c2 100644
--- a/controller/channel-billing.go
+++ b/controller/channel-billing.go
@@ -4,17 +4,18 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/monitor"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 	"strconv"
 	"time"
 
+	"github.com/songquanpeng/one-api/common/client"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/monitor"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+
 	"github.com/gin-gonic/gin"
 )
 
@@ -81,6 +82,36 @@ type APGC2DGPTUsageResponse struct {
 	TotalUsed      float64 `json:"total_used"`
 }
 
+type SiliconFlowUsageResponse struct {
+	Code    int    `json:"code"`
+	Message string `json:"message"`
+	Status  bool   `json:"status"`
+	Data    struct {
+		ID            string `json:"id"`
+		Name          string `json:"name"`
+		Image         string `json:"image"`
+		Email         string `json:"email"`
+		IsAdmin       bool   `json:"isAdmin"`
+		Balance       string `json:"balance"`
+		Status        string `json:"status"`
+		Introduction  string `json:"introduction"`
+		Role          string `json:"role"`
+		ChargeBalance string `json:"chargeBalance"`
+		TotalBalance  string `json:"totalBalance"`
+		Category      string `json:"category"`
+	} `json:"data"`
+}
+
+type DeepSeekUsageResponse struct {
+	IsAvailable  bool `json:"is_available"`
+	BalanceInfos []struct {
+		Currency        string `json:"currency"`
+		TotalBalance    string `json:"total_balance"`
+		GrantedBalance  string `json:"granted_balance"`
+		ToppedUpBalance string `json:"topped_up_balance"`
+	} `json:"balance_infos"`
+}
+
 // GetAuthHeader get auth header
 func GetAuthHeader(token string) http.Header {
 	h := http.Header{}
@@ -96,7 +127,7 @@ func GetResponseBody(method, url string, channel *model.Channel, headers http.He
 	for k := range headers {
 		req.Header.Add(k, headers.Get(k))
 	}
-	res, err := util.HTTPClient.Do(req)
+	res, err := client.HTTPClient.Do(req)
 	if err != nil {
 		return nil, err
 	}
@@ -203,30 +234,85 @@ func updateChannelAIGC2DBalance(channel *model.Channel) (float64, error) {
 	return response.TotalAvailable, nil
 }
 
+func updateChannelSiliconFlowBalance(channel *model.Channel) (float64, error) {
+	url := "https://api.siliconflow.cn/v1/user/info"
+	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
+	if err != nil {
+		return 0, err
+	}
+	response := SiliconFlowUsageResponse{}
+	err = json.Unmarshal(body, &response)
+	if err != nil {
+		return 0, err
+	}
+	if response.Code != 20000 {
+		return 0, fmt.Errorf("code: %d, message: %s", response.Code, response.Message)
+	}
+	balance, err := strconv.ParseFloat(response.Data.TotalBalance, 64)
+	if err != nil {
+		return 0, err
+	}
+	channel.UpdateBalance(balance)
+	return balance, nil
+}
+
+func updateChannelDeepSeekBalance(channel *model.Channel) (float64, error) {
+	url := "https://api.deepseek.com/user/balance"
+	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
+	if err != nil {
+		return 0, err
+	}
+	response := DeepSeekUsageResponse{}
+	err = json.Unmarshal(body, &response)
+	if err != nil {
+		return 0, err
+	}
+	index := -1
+	for i, balanceInfo := range response.BalanceInfos {
+		if balanceInfo.Currency == "CNY" {
+			index = i
+			break
+		}
+	}
+	if index == -1 {
+		return 0, errors.New("currency CNY not found")
+	}
+	balance, err := strconv.ParseFloat(response.BalanceInfos[index].TotalBalance, 64)
+	if err != nil {
+		return 0, err
+	}
+	channel.UpdateBalance(balance)
+	return balance, nil
+}
+
 func updateChannelBalance(channel *model.Channel) (float64, error) {
-	baseURL := common.ChannelBaseURLs[channel.Type]
+	baseURL := channeltype.ChannelBaseURLs[channel.Type]
 	if channel.GetBaseURL() == "" {
 		channel.BaseURL = &baseURL
 	}
 	switch channel.Type {
-	case common.ChannelTypeOpenAI:
+	case channeltype.OpenAI:
 		if channel.GetBaseURL() != "" {
 			baseURL = channel.GetBaseURL()
 		}
-	case common.ChannelTypeAzure:
+	case channeltype.Azure:
 		return 0, errors.New("尚未实现")
-	case common.ChannelTypeCustom:
+	case channeltype.Custom:
 		baseURL = channel.GetBaseURL()
-	case common.ChannelTypeCloseAI:
+	case channeltype.CloseAI:
 		return updateChannelCloseAIBalance(channel)
-	case common.ChannelTypeOpenAISB:
+	case channeltype.OpenAISB:
 		return updateChannelOpenAISBBalance(channel)
-	case common.ChannelTypeAIProxy:
+	case channeltype.AIProxy:
 		return updateChannelAIProxyBalance(channel)
-	case common.ChannelTypeAPI2GPT:
+	case channeltype.API2GPT:
 		return updateChannelAPI2GPTBalance(channel)
-	case common.ChannelTypeAIGC2D:
+	case channeltype.AIGC2D:
 		return updateChannelAIGC2DBalance(channel)
+	case channeltype.SiliconFlow:
+		return updateChannelSiliconFlowBalance(channel)
+	case channeltype.DeepSeek:
+		return updateChannelDeepSeekBalance(channel)
 	default:
 		return 0, errors.New("尚未实现")
 	}
@@ -301,11 +387,11 @@ func updateAllChannelsBalance() error {
 		return err
 	}
 	for _, channel := range channels {
-		if channel.Status != common.ChannelStatusEnabled {
+		if channel.Status != model.ChannelStatusEnabled {
 			continue
 		}
 		// TODO: support Azure
-		if channel.Type != common.ChannelTypeOpenAI && channel.Type != common.ChannelTypeCustom {
+		if channel.Type != channeltype.OpenAI && channel.Type != channeltype.Custom {
 			continue
 		}
 		balance, err := updateChannelBalance(channel)
diff --git a/controller/channel-test.go b/controller/channel-test.go
index 67ac91d0..971f5382 100644
--- a/controller/channel-test.go
+++ b/controller/channel-test.go
@@ -5,17 +5,6 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/common/message"
-	"github.com/songquanpeng/one-api/middleware"
-	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/monitor"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/helper"
-	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 	"net/http/httptest"
@@ -26,13 +15,28 @@ import (
 	"time"
 
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/message"
+	"github.com/songquanpeng/one-api/middleware"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/monitor"
+	relay "github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/controller"
+	"github.com/songquanpeng/one-api/relay/meta"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 )
 
-func buildTestRequest() *relaymodel.GeneralOpenAIRequest {
+func buildTestRequest(model string) *relaymodel.GeneralOpenAIRequest {
+	if model == "" {
+		model = "gpt-3.5-turbo"
+	}
 	testRequest := &relaymodel.GeneralOpenAIRequest{
 		MaxTokens: 2,
-		Stream:    false,
-		Model:     "gpt-3.5-turbo",
+		Model:     model,
 	}
 	testMessage := relaymodel.Message{
 		Role:    "user",
@@ -42,7 +46,7 @@ func buildTestRequest() *relaymodel.GeneralOpenAIRequest {
 	return testRequest
 }
 
-func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error) {
+func testChannel(channel *model.Channel, request *relaymodel.GeneralOpenAIRequest) (err error, openaiErr *relaymodel.Error) {
 	w := httptest.NewRecorder()
 	c, _ := gin.CreateTestContext(w)
 	c.Request = &http.Request{
@@ -53,27 +57,32 @@ func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error
 	}
 	c.Request.Header.Set("Authorization", "Bearer "+channel.Key)
 	c.Request.Header.Set("Content-Type", "application/json")
-	c.Set("channel", channel.Type)
-	c.Set("base_url", channel.GetBaseURL())
+	c.Set(ctxkey.Channel, channel.Type)
+	c.Set(ctxkey.BaseURL, channel.GetBaseURL())
+	cfg, _ := channel.LoadConfig()
+	c.Set(ctxkey.Config, cfg)
 	middleware.SetupContextForSelectedChannel(c, channel, "")
-	meta := util.GetRelayMeta(c)
-	apiType := constant.ChannelType2APIType(channel.Type)
-	adaptor := helper.GetAdaptor(apiType)
+	meta := meta.GetByContext(c)
+	apiType := channeltype.ToAPIType(channel.Type)
+	adaptor := relay.GetAdaptor(apiType)
 	if adaptor == nil {
 		return fmt.Errorf("invalid api type: %d, adaptor is nil", apiType), nil
 	}
 	adaptor.Init(meta)
-	modelName := adaptor.GetModelList()[0]
-	if !strings.Contains(channel.Models, modelName) {
+	modelName := request.Model
+	modelMap := channel.GetModelMapping()
+	if modelName == "" || !strings.Contains(channel.Models, modelName) {
 		modelNames := strings.Split(channel.Models, ",")
 		if len(modelNames) > 0 {
 			modelName = modelNames[0]
 		}
 	}
-	request := buildTestRequest()
+	if modelMap != nil && modelMap[modelName] != "" {
+		modelName = modelMap[modelName]
+	}
+	meta.OriginModelName, meta.ActualModelName = request.Model, modelName
 	request.Model = modelName
-	meta.OriginModelName, meta.ActualModelName = modelName, modelName
-	convertedRequest, err := adaptor.ConvertRequest(c, constant.RelayModeChatCompletions, request)
+	convertedRequest, err := adaptor.ConvertRequest(c, relaymode.ChatCompletions, request)
 	if err != nil {
 		return err, nil
 	}
@@ -81,14 +90,15 @@ func testChannel(channel *model.Channel) (err error, openaiErr *relaymodel.Error
 	if err != nil {
 		return err, nil
 	}
+	logger.SysLog(string(jsonData))
 	requestBody := bytes.NewBuffer(jsonData)
 	c.Request.Body = io.NopCloser(requestBody)
 	resp, err := adaptor.DoRequest(c, meta, requestBody)
 	if err != nil {
 		return err, nil
 	}
-	if resp.StatusCode != http.StatusOK {
-		err := util.RelayErrorHandler(resp)
+	if resp != nil && resp.StatusCode != http.StatusOK {
+		err := controller.RelayErrorHandler(resp)
 		return fmt.Errorf("status code %d: %s", resp.StatusCode, err.Error.Message), &err.Error
 	}
 	usage, respErr := adaptor.DoResponse(c, resp, meta)
@@ -125,10 +135,15 @@ func TestChannel(c *gin.Context) {
 		})
 		return
 	}
+	model := c.Query("model")
+	testRequest := buildTestRequest(model)
 	tik := time.Now()
-	err, _ = testChannel(channel)
+	err, _ = testChannel(channel, testRequest)
 	tok := time.Now()
 	milliseconds := tok.Sub(tik).Milliseconds()
+	if err != nil {
+		milliseconds = 0
+	}
 	go channel.UpdateResponseTime(milliseconds)
 	consumedTime := float64(milliseconds) / 1000.0
 	if err != nil {
@@ -136,6 +151,7 @@ func TestChannel(c *gin.Context) {
 			"success": false,
 			"message": err.Error(),
 			"time":    consumedTime,
+			"model":   model,
 		})
 		return
 	}
@@ -143,6 +159,7 @@ func TestChannel(c *gin.Context) {
 		"success": true,
 		"message": "",
 		"time":    consumedTime,
+		"model":   model,
 	})
 	return
 }
@@ -171,23 +188,24 @@ func testChannels(notify bool, scope string) error {
 	}
 	go func() {
 		for _, channel := range channels {
-			isChannelEnabled := channel.Status == common.ChannelStatusEnabled
+			isChannelEnabled := channel.Status == model.ChannelStatusEnabled
 			tik := time.Now()
-			err, openaiErr := testChannel(channel)
+			testRequest := buildTestRequest("")
+			err, openaiErr := testChannel(channel, testRequest)
 			tok := time.Now()
 			milliseconds := tok.Sub(tik).Milliseconds()
 			if isChannelEnabled && milliseconds > disableThreshold {
-				err = errors.New(fmt.Sprintf("响应时间 %.2fs 超过阈值 %.2fs", float64(milliseconds)/1000.0, float64(disableThreshold)/1000.0))
+				err = fmt.Errorf("响应时间 %.2fs 超过阈值 %.2fs", float64(milliseconds)/1000.0, float64(disableThreshold)/1000.0)
 				if config.AutomaticDisableChannelEnabled {
 					monitor.DisableChannel(channel.Id, channel.Name, err.Error())
 				} else {
 					_ = message.Notify(message.ByAll, fmt.Sprintf("渠道 %s （%d）测试超时", channel.Name, channel.Id), "", err.Error())
 				}
 			}
-			if isChannelEnabled && util.ShouldDisableChannel(openaiErr, -1) {
+			if isChannelEnabled && monitor.ShouldDisableChannel(openaiErr, -1) {
 				monitor.DisableChannel(channel.Id, channel.Name, err.Error())
 			}
-			if !isChannelEnabled && util.ShouldEnableChannel(err, openaiErr) {
+			if !isChannelEnabled && monitor.ShouldEnableChannel(err, openaiErr) {
 				monitor.EnableChannel(channel.Id, channel.Name)
 			}
 			channel.UpdateResponseTime(milliseconds)
@@ -197,7 +215,7 @@ func testChannels(notify bool, scope string) error {
 		testAllChannelsRunning = false
 		testAllChannelsLock.Unlock()
 		if notify {
-			err := message.Notify(message.ByAll, "通道测试完成", "", "通道测试完成，如果没有收到禁用通知，说明所有通道都正常")
+			err := message.Notify(message.ByAll, "渠道测试完成", "", "渠道测试完成，如果没有收到禁用通知，说明所有渠道都正常")
 			if err != nil {
 				logger.SysError(fmt.Sprintf("failed to send email: %s", err.Error()))
 			}
diff --git a/controller/group.go b/controller/group.go
index 128a3527..6f02394f 100644
--- a/controller/group.go
+++ b/controller/group.go
@@ -2,13 +2,13 @@ package controller
 
 import (
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
 	"net/http"
 )
 
 func GetGroups(c *gin.Context) {
 	groupNames := make([]string, 0)
-	for groupName := range common.GroupRatio {
+	for groupName := range billingratio.GroupRatio {
 		groupNames = append(groupNames, groupName)
 	}
 	c.JSON(http.StatusOK, gin.H{
diff --git a/controller/log.go b/controller/log.go
index 9377b338..665f49be 100644
--- a/controller/log.go
+++ b/controller/log.go
@@ -3,6 +3,7 @@ package controller
 import (
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
@@ -41,7 +42,7 @@ func GetUserLogs(c *gin.Context) {
 	if p < 0 {
 		p = 0
 	}
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	logType, _ := strconv.Atoi(c.Query("type"))
 	startTimestamp, _ := strconv.ParseInt(c.Query("start_timestamp"), 10, 64)
 	endTimestamp, _ := strconv.ParseInt(c.Query("end_timestamp"), 10, 64)
@@ -83,7 +84,7 @@ func SearchAllLogs(c *gin.Context) {
 
 func SearchUserLogs(c *gin.Context) {
 	keyword := c.Query("keyword")
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	logs, err := model.SearchUserLogs(userId, keyword)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -122,7 +123,7 @@ func GetLogsStat(c *gin.Context) {
 }
 
 func GetLogsSelfStat(c *gin.Context) {
-	username := c.GetString("username")
+	username := c.GetString(ctxkey.Username)
 	logType, _ := strconv.Atoi(c.Query("type"))
 	startTimestamp, _ := strconv.ParseInt(c.Query("start_timestamp"), 10, 64)
 	endTimestamp, _ := strconv.ParseInt(c.Query("end_timestamp"), 10, 64)
diff --git a/controller/misc.go b/controller/misc.go
index f27fdb12..ae900870 100644
--- a/controller/misc.go
+++ b/controller/misc.go
@@ -18,23 +18,30 @@ func GetStatus(c *gin.Context) {
 		"success": true,
 		"message": "",
 		"data": gin.H{
-			"version":             common.Version,
-			"start_time":          common.StartTime,
-			"email_verification":  config.EmailVerificationEnabled,
-			"github_oauth":        config.GitHubOAuthEnabled,
-			"github_client_id":    config.GitHubClientId,
-			"system_name":         config.SystemName,
-			"logo":                config.Logo,
-			"footer_html":         config.Footer,
-			"wechat_qrcode":       config.WeChatAccountQRCodeImageURL,
-			"wechat_login":        config.WeChatAuthEnabled,
-			"server_address":      config.ServerAddress,
-			"turnstile_check":     config.TurnstileCheckEnabled,
-			"turnstile_site_key":  config.TurnstileSiteKey,
-			"top_up_link":         config.TopUpLink,
-			"chat_link":           config.ChatLink,
-			"quota_per_unit":      config.QuotaPerUnit,
-			"display_in_currency": config.DisplayInCurrencyEnabled,
+			"version":                     common.Version,
+			"start_time":                  common.StartTime,
+			"email_verification":          config.EmailVerificationEnabled,
+			"github_oauth":                config.GitHubOAuthEnabled,
+			"github_client_id":            config.GitHubClientId,
+			"lark_client_id":              config.LarkClientId,
+			"system_name":                 config.SystemName,
+			"logo":                        config.Logo,
+			"footer_html":                 config.Footer,
+			"wechat_qrcode":               config.WeChatAccountQRCodeImageURL,
+			"wechat_login":                config.WeChatAuthEnabled,
+			"server_address":              config.ServerAddress,
+			"turnstile_check":             config.TurnstileCheckEnabled,
+			"turnstile_site_key":          config.TurnstileSiteKey,
+			"top_up_link":                 config.TopUpLink,
+			"chat_link":                   config.ChatLink,
+			"quota_per_unit":              config.QuotaPerUnit,
+			"display_in_currency":         config.DisplayInCurrencyEnabled,
+			"oidc":                        config.OidcEnabled,
+			"oidc_client_id":              config.OidcClientId,
+			"oidc_well_known":             config.OidcWellKnown,
+			"oidc_authorization_endpoint": config.OidcAuthorizationEndpoint,
+			"oidc_token_endpoint":         config.OidcTokenEndpoint,
+			"oidc_userinfo_endpoint":      config.OidcUserinfoEndpoint,
 		},
 	})
 	return
diff --git a/controller/model.go b/controller/model.go
index 4c5476b4..dcbe709e 100644
--- a/controller/model.go
+++ b/controller/model.go
@@ -3,13 +3,16 @@ package controller
 import (
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/helper"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/model"
+	relay "github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/apitype"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"net/http"
+	"strings"
 )
 
 // https://platform.openai.com/docs/api-reference/models/list
@@ -39,8 +42,8 @@ type OpenAIModels struct {
 	Parent     *string                 `json:"parent"`
 }
 
-var openAIModels []OpenAIModels
-var openAIModelsMap map[string]OpenAIModels
+var models []OpenAIModels
+var modelsMap map[string]OpenAIModels
 var channelId2Models map[int][]string
 
 func init() {
@@ -60,15 +63,15 @@ func init() {
 		IsBlocking:         false,
 	})
 	// https://platform.openai.com/docs/models/model-endpoint-compatibility
-	for i := 0; i < constant.APITypeDummy; i++ {
-		if i == constant.APITypeAIProxyLibrary {
+	for i := 0; i < apitype.Dummy; i++ {
+		if i == apitype.AIProxyLibrary {
 			continue
 		}
-		adaptor := helper.GetAdaptor(i)
+		adaptor := relay.GetAdaptor(i)
 		channelName := adaptor.GetChannelName()
 		modelNames := adaptor.GetModelList()
 		for _, modelName := range modelNames {
-			openAIModels = append(openAIModels, OpenAIModels{
+			models = append(models, OpenAIModels{
 				Id:         modelName,
 				Object:     "model",
 				Created:    1626777600,
@@ -80,12 +83,12 @@ func init() {
 		}
 	}
 	for _, channelType := range openai.CompatibleChannels {
-		if channelType == common.ChannelTypeAzure {
+		if channelType == channeltype.Azure {
 			continue
 		}
 		channelName, channelModelList := openai.GetCompatibleChannelMeta(channelType)
 		for _, modelName := range channelModelList {
-			openAIModels = append(openAIModels, OpenAIModels{
+			models = append(models, OpenAIModels{
 				Id:         modelName,
 				Object:     "model",
 				Created:    1626777600,
@@ -96,14 +99,14 @@ func init() {
 			})
 		}
 	}
-	openAIModelsMap = make(map[string]OpenAIModels)
-	for _, model := range openAIModels {
-		openAIModelsMap[model.Id] = model
+	modelsMap = make(map[string]OpenAIModels)
+	for _, model := range models {
+		modelsMap[model.Id] = model
 	}
 	channelId2Models = make(map[int][]string)
-	for i := 1; i < common.ChannelTypeDummy; i++ {
-		adaptor := helper.GetAdaptor(constant.ChannelType2APIType(i))
-		meta := &util.RelayMeta{
+	for i := 1; i < channeltype.Dummy; i++ {
+		adaptor := relay.GetAdaptor(channeltype.ToAPIType(i))
+		meta := &meta.Meta{
 			ChannelType: i,
 		}
 		adaptor.Init(meta)
@@ -119,16 +122,55 @@ func DashboardListModels(c *gin.Context) {
 	})
 }
 
-func ListModels(c *gin.Context) {
+func ListAllModels(c *gin.Context) {
 	c.JSON(200, gin.H{
 		"object": "list",
-		"data":   openAIModels,
+		"data":   models,
+	})
+}
+
+func ListModels(c *gin.Context) {
+	ctx := c.Request.Context()
+	var availableModels []string
+	if c.GetString(ctxkey.AvailableModels) != "" {
+		availableModels = strings.Split(c.GetString(ctxkey.AvailableModels), ",")
+	} else {
+		userId := c.GetInt(ctxkey.Id)
+		userGroup, _ := model.CacheGetUserGroup(userId)
+		availableModels, _ = model.CacheGetGroupModels(ctx, userGroup)
+	}
+	modelSet := make(map[string]bool)
+	for _, availableModel := range availableModels {
+		modelSet[availableModel] = true
+	}
+	availableOpenAIModels := make([]OpenAIModels, 0)
+	for _, model := range models {
+		if _, ok := modelSet[model.Id]; ok {
+			modelSet[model.Id] = false
+			availableOpenAIModels = append(availableOpenAIModels, model)
+		}
+	}
+	for modelName, ok := range modelSet {
+		if ok {
+			availableOpenAIModels = append(availableOpenAIModels, OpenAIModels{
+				Id:      modelName,
+				Object:  "model",
+				Created: 1626777600,
+				OwnedBy: "custom",
+				Root:    modelName,
+				Parent:  nil,
+			})
+		}
+	}
+	c.JSON(200, gin.H{
+		"object": "list",
+		"data":   availableOpenAIModels,
 	})
 }
 
 func RetrieveModel(c *gin.Context) {
 	modelId := c.Param("model")
-	if model, ok := openAIModelsMap[modelId]; ok {
+	if model, ok := modelsMap[modelId]; ok {
 		c.JSON(200, model)
 	} else {
 		Error := relaymodel.Error{
@@ -142,3 +184,30 @@ func RetrieveModel(c *gin.Context) {
 		})
 	}
 }
+
+func GetUserAvailableModels(c *gin.Context) {
+	ctx := c.Request.Context()
+	id := c.GetInt(ctxkey.Id)
+	userGroup, err := model.CacheGetUserGroup(id)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	models, err := model.CacheGetGroupModels(ctx, userGroup)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    models,
+	})
+	return
+}
diff --git a/controller/redemption.go b/controller/redemption.go
index 31c9348d..1d0ffbad 100644
--- a/controller/redemption.go
+++ b/controller/redemption.go
@@ -3,7 +3,9 @@ package controller
 import (
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/random"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
@@ -106,9 +108,9 @@ func AddRedemption(c *gin.Context) {
 	}
 	var keys []string
 	for i := 0; i < redemption.Count; i++ {
-		key := helper.GetUUID()
+		key := random.GetUUID()
 		cleanRedemption := model.Redemption{
-			UserId:      c.GetInt("id"),
+			UserId:      c.GetInt(ctxkey.Id),
 			Name:        redemption.Name,
 			Key:         key,
 			CreatedTime: helper.GetTimestamp(),
diff --git a/controller/relay.go b/controller/relay.go
index b34768df..038123b3 100644
--- a/controller/relay.go
+++ b/controller/relay.go
@@ -4,35 +4,38 @@ import (
 	"bytes"
 	"context"
 	"fmt"
+	"io"
+	"net/http"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/middleware"
 	dbmodel "github.com/songquanpeng/one-api/model"
 	"github.com/songquanpeng/one-api/monitor"
-	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/controller"
 	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 )
 
 // https://platform.openai.com/docs/api-reference/chat
 
-func relay(c *gin.Context, relayMode int) *model.ErrorWithStatusCode {
+func relayHelper(c *gin.Context, relayMode int) *model.ErrorWithStatusCode {
 	var err *model.ErrorWithStatusCode
 	switch relayMode {
-	case constant.RelayModeImagesGenerations:
+	case relaymode.ImagesGenerations:
 		err = controller.RelayImageHelper(c, relayMode)
-	case constant.RelayModeAudioSpeech:
+	case relaymode.AudioSpeech:
 		fallthrough
-	case constant.RelayModeAudioTranslation:
+	case relaymode.AudioTranslation:
 		fallthrough
-	case constant.RelayModeAudioTranscription:
+	case relaymode.AudioTranscription:
 		err = controller.RelayAudioHelper(c, relayMode)
+	case relaymode.Proxy:
+		err = controller.RelayProxyHelper(c, relayMode)
 	default:
 		err = controller.RelayTextHelper(c)
 	}
@@ -41,23 +44,24 @@ func relay(c *gin.Context, relayMode int) *model.ErrorWithStatusCode {
 
 func Relay(c *gin.Context) {
 	ctx := c.Request.Context()
-	relayMode := constant.Path2RelayMode(c.Request.URL.Path)
+	relayMode := relaymode.GetByPath(c.Request.URL.Path)
 	if config.DebugEnabled {
 		requestBody, _ := common.GetRequestBody(c)
 		logger.Debugf(ctx, "request body: %s", string(requestBody))
 	}
-	channelId := c.GetInt("channel_id")
-	bizErr := relay(c, relayMode)
+	channelId := c.GetInt(ctxkey.ChannelId)
+	userId := c.GetInt(ctxkey.Id)
+	bizErr := relayHelper(c, relayMode)
 	if bizErr == nil {
 		monitor.Emit(channelId, true)
 		return
 	}
 	lastFailedChannelId := channelId
-	channelName := c.GetString("channel_name")
-	group := c.GetString("group")
-	originalModel := c.GetString("original_model")
-	go processChannelRelayError(ctx, channelId, channelName, bizErr)
-	requestId := c.GetString(logger.RequestIdKey)
+	channelName := c.GetString(ctxkey.ChannelName)
+	group := c.GetString(ctxkey.Group)
+	originalModel := c.GetString(ctxkey.OriginalModel)
+	go processChannelRelayError(ctx, userId, channelId, channelName, *bizErr)
+	requestId := c.GetString(helper.RequestIdKey)
 	retryTimes := config.RetryTimes
 	if !shouldRetry(c, bizErr.StatusCode) {
 		logger.Errorf(ctx, "relay error happen, status code is %d, won't retry in this case", bizErr.StatusCode)
@@ -66,7 +70,7 @@ func Relay(c *gin.Context) {
 	for i := retryTimes; i > 0; i-- {
 		channel, err := dbmodel.CacheGetRandomSatisfiedChannel(group, originalModel, i != retryTimes)
 		if err != nil {
-			logger.Errorf(ctx, "CacheGetRandomSatisfiedChannel failed: %w", err)
+			logger.Errorf(ctx, "CacheGetRandomSatisfiedChannel failed: %+v", err)
 			break
 		}
 		logger.Infof(ctx, "using channel #%d to retry (remain times %d)", channel.Id, i)
@@ -76,19 +80,21 @@ func Relay(c *gin.Context) {
 		middleware.SetupContextForSelectedChannel(c, channel, originalModel)
 		requestBody, err := common.GetRequestBody(c)
 		c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
-		bizErr = relay(c, relayMode)
+		bizErr = relayHelper(c, relayMode)
 		if bizErr == nil {
 			return
 		}
-		channelId := c.GetInt("channel_id")
+		channelId := c.GetInt(ctxkey.ChannelId)
 		lastFailedChannelId = channelId
-		channelName := c.GetString("channel_name")
-		go processChannelRelayError(ctx, channelId, channelName, bizErr)
+		channelName := c.GetString(ctxkey.ChannelName)
+		go processChannelRelayError(ctx, userId, channelId, channelName, *bizErr)
 	}
 	if bizErr != nil {
 		if bizErr.StatusCode == http.StatusTooManyRequests {
 			bizErr.Error.Message = "当前分组上游负载已饱和，请稍后再试"
 		}
+
+		// BUG: bizErr is in race condition
 		bizErr.Error.Message = helper.MessageWithRequestId(bizErr.Error.Message, requestId)
 		c.JSON(bizErr.StatusCode, gin.H{
 			"error": bizErr.Error,
@@ -97,7 +103,7 @@ func Relay(c *gin.Context) {
 }
 
 func shouldRetry(c *gin.Context, statusCode int) bool {
-	if _, ok := c.Get("specific_channel_id"); ok {
+	if _, ok := c.Get(ctxkey.SpecificChannelId); ok {
 		return false
 	}
 	if statusCode == http.StatusTooManyRequests {
@@ -115,10 +121,10 @@ func shouldRetry(c *gin.Context, statusCode int) bool {
 	return true
 }
 
-func processChannelRelayError(ctx context.Context, channelId int, channelName string, err *model.ErrorWithStatusCode) {
-	logger.Errorf(ctx, "relay error (channel #%d): %s", channelId, err.Message)
+func processChannelRelayError(ctx context.Context, userId int, channelId int, channelName string, err model.ErrorWithStatusCode) {
+	logger.Errorf(ctx, "relay error (channel id %d, user id: %d): %s", channelId, userId, err.Message)
 	// https://platform.openai.com/docs/guides/error-codes/api-errors
-	if util.ShouldDisableChannel(&err.Error, err.StatusCode) {
+	if monitor.ShouldDisableChannel(&err.Error, err.StatusCode) {
 		monitor.DisableChannel(channelId, channelName, err.Message)
 	} else {
 		monitor.Emit(channelId, false)
diff --git a/controller/token.go b/controller/token.go
index 7f6b4505..668ccd97 100644
--- a/controller/token.go
+++ b/controller/token.go
@@ -1,17 +1,20 @@
 package controller
 
 import (
+	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/network"
+	"github.com/songquanpeng/one-api/common/random"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
 )
 
 func GetAllTokens(c *gin.Context) {
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	p, _ := strconv.Atoi(c.Query("p"))
 	if p < 0 {
 		p = 0
@@ -36,7 +39,7 @@ func GetAllTokens(c *gin.Context) {
 }
 
 func SearchTokens(c *gin.Context) {
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	keyword := c.Query("keyword")
 	tokens, err := model.SearchUserTokens(userId, keyword)
 	if err != nil {
@@ -56,7 +59,7 @@ func SearchTokens(c *gin.Context) {
 
 func GetToken(c *gin.Context) {
 	id, err := strconv.Atoi(c.Param("id"))
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
@@ -81,8 +84,8 @@ func GetToken(c *gin.Context) {
 }
 
 func GetTokenStatus(c *gin.Context) {
-	tokenId := c.GetInt("token_id")
-	userId := c.GetInt("id")
+	tokenId := c.GetInt(ctxkey.TokenId)
+	userId := c.GetInt(ctxkey.Id)
 	token, err := model.GetTokenByIds(tokenId, userId)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -104,6 +107,19 @@ func GetTokenStatus(c *gin.Context) {
 	})
 }
 
+func validateToken(c *gin.Context, token model.Token) error {
+	if len(token.Name) > 30 {
+		return fmt.Errorf("令牌名称过长")
+	}
+	if token.Subnet != nil && *token.Subnet != "" {
+		err := network.IsValidSubnets(*token.Subnet)
+		if err != nil {
+			return fmt.Errorf("无效的网段：%s", err.Error())
+		}
+	}
+	return nil
+}
+
 func AddToken(c *gin.Context) {
 	token := model.Token{}
 	err := c.ShouldBindJSON(&token)
@@ -114,22 +130,26 @@ func AddToken(c *gin.Context) {
 		})
 		return
 	}
-	if len(token.Name) > 30 {
+	err = validateToken(c, token)
+	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
-			"message": "令牌名称过长",
+			"message": fmt.Sprintf("参数错误：%s", err.Error()),
 		})
 		return
 	}
+
 	cleanToken := model.Token{
-		UserId:         c.GetInt("id"),
+		UserId:         c.GetInt(ctxkey.Id),
 		Name:           token.Name,
-		Key:            helper.GenerateKey(),
+		Key:            random.GenerateKey(),
 		CreatedTime:    helper.GetTimestamp(),
 		AccessedTime:   helper.GetTimestamp(),
 		ExpiredTime:    token.ExpiredTime,
 		RemainQuota:    token.RemainQuota,
 		UnlimitedQuota: token.UnlimitedQuota,
+		Models:         token.Models,
+		Subnet:         token.Subnet,
 	}
 	err = cleanToken.Insert()
 	if err != nil {
@@ -142,13 +162,14 @@ func AddToken(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{
 		"success": true,
 		"message": "",
+		"data":    cleanToken,
 	})
 	return
 }
 
 func DeleteToken(c *gin.Context) {
 	id, _ := strconv.Atoi(c.Param("id"))
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	err := model.DeleteTokenById(id, userId)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -165,7 +186,7 @@ func DeleteToken(c *gin.Context) {
 }
 
 func UpdateToken(c *gin.Context) {
-	userId := c.GetInt("id")
+	userId := c.GetInt(ctxkey.Id)
 	statusOnly := c.Query("status_only")
 	token := model.Token{}
 	err := c.ShouldBindJSON(&token)
@@ -176,10 +197,11 @@ func UpdateToken(c *gin.Context) {
 		})
 		return
 	}
-	if len(token.Name) > 30 {
+	err = validateToken(c, token)
+	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
-			"message": "令牌名称过长",
+			"message": fmt.Sprintf("参数错误：%s", err.Error()),
 		})
 		return
 	}
@@ -191,15 +213,15 @@ func UpdateToken(c *gin.Context) {
 		})
 		return
 	}
-	if token.Status == common.TokenStatusEnabled {
-		if cleanToken.Status == common.TokenStatusExpired && cleanToken.ExpiredTime <= helper.GetTimestamp() && cleanToken.ExpiredTime != -1 {
+	if token.Status == model.TokenStatusEnabled {
+		if cleanToken.Status == model.TokenStatusExpired && cleanToken.ExpiredTime <= helper.GetTimestamp() && cleanToken.ExpiredTime != -1 {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "令牌已过期，无法启用，请先修改令牌过期时间，或者设置为永不过期",
 			})
 			return
 		}
-		if cleanToken.Status == common.TokenStatusExhausted && cleanToken.RemainQuota <= 0 && !cleanToken.UnlimitedQuota {
+		if cleanToken.Status == model.TokenStatusExhausted && cleanToken.RemainQuota <= 0 && !cleanToken.UnlimitedQuota {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "令牌可用额度已用尽，无法启用，请先修改令牌剩余额度，或者设置为无限额度",
@@ -215,6 +237,8 @@ func UpdateToken(c *gin.Context) {
 		cleanToken.ExpiredTime = token.ExpiredTime
 		cleanToken.RemainQuota = token.RemainQuota
 		cleanToken.UnlimitedQuota = token.UnlimitedQuota
+		cleanToken.Models = token.Models
+		cleanToken.Subnet = token.Subnet
 	}
 	err = cleanToken.Update()
 	if err != nil {
diff --git a/controller/user.go b/controller/user.go
index 8b614e5d..e79881c2 100644
--- a/controller/user.go
+++ b/controller/user.go
@@ -5,7 +5,8 @@ import (
 	"fmt"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/random"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strconv"
@@ -58,11 +59,11 @@ func Login(c *gin.Context) {
 		})
 		return
 	}
-	setupLogin(&user, c)
+	SetupLogin(&user, c)
 }
 
 // setup session & cookies and then return user info
-func setupLogin(user *model.User, c *gin.Context) {
+func SetupLogin(user *model.User, c *gin.Context) {
 	session := sessions.Default(c)
 	session.Set("id", user.Id)
 	session.Set("username", user.Username)
@@ -172,6 +173,7 @@ func Register(c *gin.Context) {
 		})
 		return
 	}
+
 	c.JSON(http.StatusOK, gin.H{
 		"success": true,
 		"message": "",
@@ -180,27 +182,27 @@ func Register(c *gin.Context) {
 }
 
 func GetAllUsers(c *gin.Context) {
-    p, _ := strconv.Atoi(c.Query("p"))
-    if p < 0 {
-        p = 0
-    }
-    
-    order := c.DefaultQuery("order", "")
-    users, err := model.GetAllUsers(p*config.ItemsPerPage, config.ItemsPerPage, order)
-	
-    if err != nil {
-        c.JSON(http.StatusOK, gin.H{
-            "success": false,
-            "message": err.Error(),
-        })
-        return
-    }
-    
-    c.JSON(http.StatusOK, gin.H{
-        "success": true,
-        "message": "",
-        "data":    users,
-    })
+	p, _ := strconv.Atoi(c.Query("p"))
+	if p < 0 {
+		p = 0
+	}
+
+	order := c.DefaultQuery("order", "")
+	users, err := model.GetAllUsers(p*config.ItemsPerPage, config.ItemsPerPage, order)
+
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    users,
+	})
 }
 
 func SearchUsers(c *gin.Context) {
@@ -238,8 +240,8 @@ func GetUser(c *gin.Context) {
 		})
 		return
 	}
-	myRole := c.GetInt("role")
-	if myRole <= user.Role && myRole != common.RoleRootUser {
+	myRole := c.GetInt(ctxkey.Role)
+	if myRole <= user.Role && myRole != model.RoleRootUser {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "无权获取同级或更高等级用户的信息",
@@ -255,7 +257,7 @@ func GetUser(c *gin.Context) {
 }
 
 func GetUserDashboard(c *gin.Context) {
-	id := c.GetInt("id")
+	id := c.GetInt(ctxkey.Id)
 	now := time.Now()
 	startOfDay := now.Truncate(24*time.Hour).AddDate(0, 0, -6).Unix()
 	endOfDay := now.Truncate(24 * time.Hour).Add(24*time.Hour - time.Second).Unix()
@@ -278,7 +280,7 @@ func GetUserDashboard(c *gin.Context) {
 }
 
 func GenerateAccessToken(c *gin.Context) {
-	id := c.GetInt("id")
+	id := c.GetInt(ctxkey.Id)
 	user, err := model.GetUserById(id, true)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -287,7 +289,7 @@ func GenerateAccessToken(c *gin.Context) {
 		})
 		return
 	}
-	user.AccessToken = helper.GetUUID()
+	user.AccessToken = random.GetUUID()
 
 	if model.DB.Where("access_token = ?", user.AccessToken).First(user).RowsAffected != 0 {
 		c.JSON(http.StatusOK, gin.H{
@@ -314,7 +316,7 @@ func GenerateAccessToken(c *gin.Context) {
 }
 
 func GetAffCode(c *gin.Context) {
-	id := c.GetInt("id")
+	id := c.GetInt(ctxkey.Id)
 	user, err := model.GetUserById(id, true)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -324,7 +326,7 @@ func GetAffCode(c *gin.Context) {
 		return
 	}
 	if user.AffCode == "" {
-		user.AffCode = helper.GetRandomString(4)
+		user.AffCode = random.GetRandomString(4)
 		if err := user.Update(false); err != nil {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
@@ -342,7 +344,7 @@ func GetAffCode(c *gin.Context) {
 }
 
 func GetSelf(c *gin.Context) {
-	id := c.GetInt("id")
+	id := c.GetInt(ctxkey.Id)
 	user, err := model.GetUserById(id, false)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -387,15 +389,15 @@ func UpdateUser(c *gin.Context) {
 		})
 		return
 	}
-	myRole := c.GetInt("role")
-	if myRole <= originUser.Role && myRole != common.RoleRootUser {
+	myRole := c.GetInt(ctxkey.Role)
+	if myRole <= originUser.Role && myRole != model.RoleRootUser {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "无权更新同权限等级或更高权限等级的用户信息",
 		})
 		return
 	}
-	if myRole <= updatedUser.Role && myRole != common.RoleRootUser {
+	if myRole <= updatedUser.Role && myRole != model.RoleRootUser {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "无权将其他用户权限等级提升到大于等于自己的权限等级",
@@ -445,7 +447,7 @@ func UpdateSelf(c *gin.Context) {
 	}
 
 	cleanUser := model.User{
-		Id:          c.GetInt("id"),
+		Id:          c.GetInt(ctxkey.Id),
 		Username:    user.Username,
 		Password:    user.Password,
 		DisplayName: user.DisplayName,
@@ -509,7 +511,7 @@ func DeleteSelf(c *gin.Context) {
 	id := c.GetInt("id")
 	user, _ := model.GetUserById(id, false)
 
-	if user.Role == common.RoleRootUser {
+	if user.Role == model.RoleRootUser {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "不能删除超级管理员账户",
@@ -611,7 +613,7 @@ func ManageUser(c *gin.Context) {
 		return
 	}
 	myRole := c.GetInt("role")
-	if myRole <= user.Role && myRole != common.RoleRootUser {
+	if myRole <= user.Role && myRole != model.RoleRootUser {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "无权更新同权限等级或更高权限等级的用户信息",
@@ -620,8 +622,8 @@ func ManageUser(c *gin.Context) {
 	}
 	switch req.Action {
 	case "disable":
-		user.Status = common.UserStatusDisabled
-		if user.Role == common.RoleRootUser {
+		user.Status = model.UserStatusDisabled
+		if user.Role == model.RoleRootUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "无法禁用超级管理员用户",
@@ -629,9 +631,9 @@ func ManageUser(c *gin.Context) {
 			return
 		}
 	case "enable":
-		user.Status = common.UserStatusEnabled
+		user.Status = model.UserStatusEnabled
 	case "delete":
-		if user.Role == common.RoleRootUser {
+		if user.Role == model.RoleRootUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "无法删除超级管理员用户",
@@ -646,37 +648,37 @@ func ManageUser(c *gin.Context) {
 			return
 		}
 	case "promote":
-		if myRole != common.RoleRootUser {
+		if myRole != model.RoleRootUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "普通管理员用户无法提升其他用户为管理员",
 			})
 			return
 		}
-		if user.Role >= common.RoleAdminUser {
+		if user.Role >= model.RoleAdminUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "该用户已经是管理员",
 			})
 			return
 		}
-		user.Role = common.RoleAdminUser
+		user.Role = model.RoleAdminUser
 	case "demote":
-		if user.Role == common.RoleRootUser {
+		if user.Role == model.RoleRootUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "无法降级超级管理员用户",
 			})
 			return
 		}
-		if user.Role == common.RoleCommonUser {
+		if user.Role == model.RoleCommonUser {
 			c.JSON(http.StatusOK, gin.H{
 				"success": false,
 				"message": "该用户已经是普通用户",
 			})
 			return
 		}
-		user.Role = common.RoleCommonUser
+		user.Role = model.RoleCommonUser
 	}
 
 	if err := user.Update(false); err != nil {
@@ -730,7 +732,7 @@ func EmailBind(c *gin.Context) {
 		})
 		return
 	}
-	if user.Role == common.RoleRootUser {
+	if user.Role == model.RoleRootUser {
 		config.RootUserEmail = email
 	}
 	c.JSON(http.StatusOK, gin.H{
@@ -770,3 +772,38 @@ func TopUp(c *gin.Context) {
 	})
 	return
 }
+
+type adminTopUpRequest struct {
+	UserId int    `json:"user_id"`
+	Quota  int    `json:"quota"`
+	Remark string `json:"remark"`
+}
+
+func AdminTopUp(c *gin.Context) {
+	req := adminTopUpRequest{}
+	err := c.ShouldBindJSON(&req)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	err = model.IncreaseUserQuota(req.UserId, int64(req.Quota))
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	if req.Remark == "" {
+		req.Remark = fmt.Sprintf("通过 API 充值 %s", common.LogQuota(int64(req.Quota)))
+	}
+	model.RecordTopupLog(req.UserId, req.Remark, req.Quota)
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+	})
+	return
+}
diff --git a/docs/API.md b/docs/API.md
new file mode 100644
index 00000000..0b7ddf5a
--- /dev/null
+++ b/docs/API.md
@@ -0,0 +1,53 @@
+# 使用 API 操控 & 扩展 One API
+> 欢迎提交 PR 在此放上你的拓展项目。
+
+例如，虽然 One API 本身没有直接支持支付，但是你可以通过系统扩展的 API 来实现支付功能。
+
+又或者你想自定义渠道管理策略，也可以通过 API 来实现渠道的禁用与启用。
+
+## 鉴权
+One API 支持两种鉴权方式：Cookie 和 Token，对于 Token，参照下图获取：
+
+![image](https://github.com/songquanpeng/songquanpeng.github.io/assets/39998050/c15281a7-83ed-47cb-a1f6-913cb6bf4a7c)
+
+之后，将 Token 作为请求头的 Authorization 字段的值即可，例如下面使用 Token 调用测试渠道的 API：
+![image](https://github.com/songquanpeng/songquanpeng.github.io/assets/39998050/1273b7ae-cb60-4c0d-93a6-b1cbc039c4f8)
+
+## 请求格式与响应格式
+One API 使用 JSON 格式进行请求和响应。
+
+对于响应体，一般格式如下：
+```json
+{
+  "message": "请求信息",
+  "success": true,
+  "data": {}
+}
+```
+
+## API 列表
+> 当前 API 列表不全，请自行通过浏览器抓取前端请求
+
+如果现有的 API 没有办法满足你的需求，欢迎提交 issue 讨论。
+
+### 获取当前登录用户信息
+**GET** `/api/user/self`
+
+### 为给定用户充值额度
+**POST** `/api/topup`
+```json
+{
+  "user_id": 1,
+  "quota": 100000,
+  "remark": "充值 100000 额度"
+}
+```
+
+## 其他
+### 充值链接上的附加参数
+One API 会在用户点击充值按钮的时候，将用户的信息和充值信息附加在链接上，例如：
+`https://example.com?username=root&user_id=1&transaction_id=4b3eed80-55d5-443f-bd44-fb18c648c837`
+
+你可以通过解析链接上的参数来获取用户信息和充值信息，然后调用 API 来为用户充值。
+
+注意，不是所有主题都支持该功能，欢迎 PR 补齐。
\ No newline at end of file
diff --git a/go.mod b/go.mod
index f9ed96d3..2106cf0f 100644
--- a/go.mod
+++ b/go.mod
@@ -1,67 +1,111 @@
 module github.com/songquanpeng/one-api
 
 // +heroku goVersion go1.18
-go 1.18
+go 1.20
 
 require (
-	github.com/gin-contrib/cors v1.4.0
-	github.com/gin-contrib/gzip v0.0.6
-	github.com/gin-contrib/sessions v0.0.5
-	github.com/gin-contrib/static v0.0.1
-	github.com/gin-gonic/gin v1.9.1
-	github.com/go-playground/validator/v10 v10.14.0
+	cloud.google.com/go/iam v1.1.10
+	github.com/aws/aws-sdk-go-v2 v1.27.0
+	github.com/aws/aws-sdk-go-v2/credentials v1.17.15
+	github.com/aws/aws-sdk-go-v2/service/bedrockruntime v1.8.3
+	github.com/gin-contrib/cors v1.7.2
+	github.com/gin-contrib/gzip v1.0.1
+	github.com/gin-contrib/sessions v1.0.1
+	github.com/gin-contrib/static v1.1.2
+	github.com/gin-gonic/gin v1.10.0
+	github.com/go-playground/validator/v10 v10.20.0
 	github.com/go-redis/redis/v8 v8.11.5
 	github.com/golang-jwt/jwt v3.2.2+incompatible
-	github.com/google/uuid v1.3.0
-	github.com/gorilla/websocket v1.5.0
-	github.com/pkoukk/tiktoken-go v0.1.5
-	github.com/stretchr/testify v1.8.3
-	golang.org/x/crypto v0.17.0
-	golang.org/x/image v0.14.0
-	gorm.io/driver/mysql v1.4.3
-	gorm.io/driver/postgres v1.5.2
-	gorm.io/driver/sqlite v1.4.3
-	gorm.io/gorm v1.25.0
+	github.com/google/uuid v1.6.0
+	github.com/gorilla/websocket v1.5.1
+	github.com/jinzhu/copier v0.4.0
+	github.com/joho/godotenv v1.5.1
+	github.com/patrickmn/go-cache v2.1.0+incompatible
+	github.com/pkg/errors v0.9.1
+	github.com/pkoukk/tiktoken-go v0.1.7
+	github.com/smartystreets/goconvey v1.8.1
+	github.com/stretchr/testify v1.9.0
+	golang.org/x/crypto v0.31.0
+	golang.org/x/image v0.18.0
+	google.golang.org/api v0.187.0
+	gorm.io/driver/mysql v1.5.6
+	gorm.io/driver/postgres v1.5.7
+	gorm.io/driver/sqlite v1.5.5
+	gorm.io/gorm v1.25.10
 )
 
 require (
-	github.com/bytedance/sonic v1.9.1 // indirect
-	github.com/cespare/xxhash/v2 v2.1.2 // indirect
-	github.com/chenzhuoyu/base64x v0.0.0-20221115062448-fe3a3abad311 // indirect
+	cloud.google.com/go/auth v0.6.1 // indirect
+	cloud.google.com/go/auth/oauth2adapt v0.2.2 // indirect
+	cloud.google.com/go/compute/metadata v0.3.0 // indirect
+	filippo.io/edwards25519 v1.1.0 // indirect
+	github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.2 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.7 // indirect
+	github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.7 // indirect
+	github.com/aws/smithy-go v1.20.2 // indirect
+	github.com/bytedance/sonic v1.11.6 // indirect
+	github.com/bytedance/sonic/loader v0.1.1 // indirect
+	github.com/cespare/xxhash/v2 v2.3.0 // indirect
+	github.com/cloudwego/base64x v0.1.4 // indirect
+	github.com/cloudwego/iasm v0.2.0 // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect
-	github.com/dlclark/regexp2 v1.10.0 // indirect
-	github.com/gabriel-vasile/mimetype v1.4.2 // indirect
+	github.com/dlclark/regexp2 v1.11.0 // indirect
+	github.com/felixge/httpsnoop v1.0.4 // indirect
+	github.com/fsnotify/fsnotify v1.7.0 // indirect
+	github.com/gabriel-vasile/mimetype v1.4.3 // indirect
 	github.com/gin-contrib/sse v0.1.0 // indirect
+	github.com/go-logr/logr v1.4.1 // indirect
+	github.com/go-logr/stdr v1.2.2 // indirect
 	github.com/go-playground/locales v0.14.1 // indirect
 	github.com/go-playground/universal-translator v0.18.1 // indirect
-	github.com/go-sql-driver/mysql v1.6.0 // indirect
-	github.com/goccy/go-json v0.10.2 // indirect
-	github.com/gorilla/context v1.1.1 // indirect
-	github.com/gorilla/securecookie v1.1.1 // indirect
-	github.com/gorilla/sessions v1.2.1 // indirect
+	github.com/go-sql-driver/mysql v1.8.1 // indirect
+	github.com/goccy/go-json v0.10.3 // indirect
+	github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da // indirect
+	github.com/golang/protobuf v1.5.4 // indirect
+	github.com/google/s2a-go v0.1.7 // indirect
+	github.com/googleapis/enterprise-certificate-proxy v0.3.2 // indirect
+	github.com/googleapis/gax-go/v2 v2.12.5 // indirect
+	github.com/gopherjs/gopherjs v1.17.2 // indirect
+	github.com/gorilla/context v1.1.2 // indirect
+	github.com/gorilla/securecookie v1.1.2 // indirect
+	github.com/gorilla/sessions v1.2.2 // indirect
 	github.com/jackc/pgpassfile v1.0.0 // indirect
-	github.com/jackc/pgservicefile v0.0.0-20221227161230-091c0ba34f0a // indirect
-	github.com/jackc/pgx/v5 v5.5.4 // indirect
+	github.com/jackc/pgservicefile v0.0.0-20231201235250-de7065d80cb9 // indirect
+	github.com/jackc/pgx/v5 v5.5.5 // indirect
 	github.com/jackc/puddle/v2 v2.2.1 // indirect
 	github.com/jinzhu/inflection v1.0.0 // indirect
 	github.com/jinzhu/now v1.1.5 // indirect
 	github.com/json-iterator/go v1.1.12 // indirect
-	github.com/klauspost/cpuid/v2 v2.2.4 // indirect
-	github.com/leodido/go-urn v1.2.4 // indirect
-	github.com/mattn/go-isatty v0.0.19 // indirect
-	github.com/mattn/go-sqlite3 v2.0.3+incompatible // indirect
+	github.com/jtolds/gls v4.20.0+incompatible // indirect
+	github.com/klauspost/cpuid/v2 v2.2.7 // indirect
+	github.com/kr/text v0.2.0 // indirect
+	github.com/leodido/go-urn v1.4.0 // indirect
+	github.com/mattn/go-isatty v0.0.20 // indirect
+	github.com/mattn/go-sqlite3 v1.14.22 // indirect
 	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
 	github.com/modern-go/reflect2 v1.0.2 // indirect
-	github.com/pelletier/go-toml/v2 v2.0.8 // indirect
+	github.com/pelletier/go-toml/v2 v2.2.2 // indirect
 	github.com/pmezard/go-difflib v1.0.0 // indirect
+	github.com/smarty/assertions v1.15.0 // indirect
 	github.com/twitchyliquid64/golang-asm v0.15.1 // indirect
-	github.com/ugorji/go/codec v1.2.11 // indirect
-	golang.org/x/arch v0.3.0 // indirect
-	golang.org/x/net v0.17.0 // indirect
-	golang.org/x/sync v0.1.0 // indirect
-	golang.org/x/sys v0.15.0 // indirect
-	golang.org/x/text v0.14.0 // indirect
-	google.golang.org/protobuf v1.33.0 // indirect
+	github.com/ugorji/go/codec v1.2.12 // indirect
+	go.opencensus.io v0.24.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 // indirect
+	go.opentelemetry.io/otel v1.24.0 // indirect
+	go.opentelemetry.io/otel/metric v1.24.0 // indirect
+	go.opentelemetry.io/otel/trace v1.24.0 // indirect
+	golang.org/x/arch v0.8.0 // indirect
+	golang.org/x/net v0.26.0 // indirect
+	golang.org/x/oauth2 v0.21.0 // indirect
+	golang.org/x/sync v0.10.0 // indirect
+	golang.org/x/sys v0.28.0 // indirect
+	golang.org/x/text v0.21.0 // indirect
+	golang.org/x/time v0.5.0 // indirect
+	google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d // indirect
+	google.golang.org/grpc v1.64.1 // indirect
+	google.golang.org/protobuf v1.34.2 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // indirect
 )
diff --git a/go.sum b/go.sum
index 9cf056e5..c98f1965 100644
--- a/go.sum
+++ b/go.sum
@@ -1,210 +1,317 @@
-github.com/bytedance/sonic v1.5.0/go.mod h1:ED5hyg4y6t3/9Ku1R6dU/4KyJ48DZ4jPhfY1O2AihPM=
-github.com/bytedance/sonic v1.9.1 h1:6iJ6NqdoxCDr6mbY8h18oSO+cShGSMRGCEo7F2h0x8s=
-github.com/bytedance/sonic v1.9.1/go.mod h1:i736AoUSYt75HyZLoJW9ERYxcy6eaN6h4BZXU064P/U=
-github.com/cespare/xxhash/v2 v2.1.2 h1:YRXhKfTDauu4ajMg1TPgFO5jnlC2HCbmLXMcTG5cbYE=
-github.com/cespare/xxhash/v2 v2.1.2/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
-github.com/chenzhuoyu/base64x v0.0.0-20211019084208-fb5309c8db06/go.mod h1:DH46F32mSOjUmXrMHnKwZdA8wcEefY7UVqBKYGjpdQY=
-github.com/chenzhuoyu/base64x v0.0.0-20221115062448-fe3a3abad311 h1:qSGYFH7+jGhDF8vLC+iwCD4WpbV1EBDSzWkJODFLams=
-github.com/chenzhuoyu/base64x v0.0.0-20221115062448-fe3a3abad311/go.mod h1:b583jCggY9gE99b6G5LEC39OIiVsWj+R97kbl5odCEk=
+cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
+cloud.google.com/go/auth v0.6.1 h1:T0Zw1XM5c1GlpN2HYr2s+m3vr1p2wy+8VN+Z1FKxW38=
+cloud.google.com/go/auth v0.6.1/go.mod h1:eFHG7zDzbXHKmjJddFG/rBlcGp6t25SwRUiEQSlO4x4=
+cloud.google.com/go/auth/oauth2adapt v0.2.2 h1:+TTV8aXpjeChS9M+aTtN/TjdQnzJvmzKFt//oWu7HX4=
+cloud.google.com/go/auth/oauth2adapt v0.2.2/go.mod h1:wcYjgpZI9+Yu7LyYBg4pqSiaRkfEK3GQcpb7C/uyF1Q=
+cloud.google.com/go/compute/metadata v0.3.0 h1:Tz+eQXMEqDIKRsmY3cHTL6FVaynIjX2QxYC4trgAKZc=
+cloud.google.com/go/compute/metadata v0.3.0/go.mod h1:zFmK7XCadkQkj6TtorcaGlCW1hT1fIilQDwofLpJ20k=
+cloud.google.com/go/iam v1.1.10 h1:ZSAr64oEhQSClwBL670MsJAW5/RLiC6kfw3Bqmd5ZDI=
+cloud.google.com/go/iam v1.1.10/go.mod h1:iEgMq62sg8zx446GCaijmA2Miwg5o3UbO+nI47WHJps=
+filippo.io/edwards25519 v1.1.0 h1:FNf4tywRC1HmFuKW5xopWpigGjJKiJSV0Cqo0cJWDaA=
+filippo.io/edwards25519 v1.1.0/go.mod h1:BxyFTGdWcka3PhytdK4V28tE5sGfRvvvRV7EaN4VDT4=
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
+github.com/aws/aws-sdk-go-v2 v1.27.0 h1:7bZWKoXhzI+mMR/HjdMx8ZCC5+6fY0lS5tr0bbgiLlo=
+github.com/aws/aws-sdk-go-v2 v1.27.0/go.mod h1:ffIFB97e2yNsv4aTSGkqtHnppsIJzw7G7BReUZ3jCXM=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.2 h1:x6xsQXGSmW6frevwDA+vi/wqhp1ct18mVXYN08/93to=
+github.com/aws/aws-sdk-go-v2/aws/protocol/eventstream v1.6.2/go.mod h1:lPprDr1e6cJdyYeGXnRaJoP4Md+cDBvi2eOj00BlGmg=
+github.com/aws/aws-sdk-go-v2/credentials v1.17.15 h1:YDexlvDRCA8ems2T5IP1xkMtOZ1uLJOCJdTr0igs5zo=
+github.com/aws/aws-sdk-go-v2/credentials v1.17.15/go.mod h1:vxHggqW6hFNaeNC0WyXS3VdyjcV0a4KMUY4dKJ96buU=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.7 h1:lf/8VTF2cM+N4SLzaYJERKEWAXq8MOMpZfU6wEPWsPk=
+github.com/aws/aws-sdk-go-v2/internal/configsources v1.3.7/go.mod h1:4SjkU7QiqK2M9oozyMzfZ/23LmUY+h3oFqhdeP5OMiI=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.7 h1:4OYVp0705xu8yjdyoWix0r9wPIRXnIzzOoUpQVHIJ/g=
+github.com/aws/aws-sdk-go-v2/internal/endpoints/v2 v2.6.7/go.mod h1:vd7ESTEvI76T2Na050gODNmNU7+OyKrIKroYTu4ABiI=
+github.com/aws/aws-sdk-go-v2/service/bedrockruntime v1.8.3 h1:Fihjyd6DeNjcawBEGLH9dkIEUi6AdhucDKPE9nJ4QiY=
+github.com/aws/aws-sdk-go-v2/service/bedrockruntime v1.8.3/go.mod h1:opvUj3ismqSCxYc+m4WIjPL0ewZGtvp0ess7cKvBPOQ=
+github.com/aws/smithy-go v1.20.2 h1:tbp628ireGtzcHDDmLT/6ADHidqnwgF57XOXZe6tp4Q=
+github.com/aws/smithy-go v1.20.2/go.mod h1:krry+ya/rV9RDcV/Q16kpu6ypI4K2czasz0NC3qS14E=
+github.com/bytedance/sonic v1.11.6 h1:oUp34TzMlL+OY1OUWxHqsdkgC/Zfc85zGqw9siXjrc0=
+github.com/bytedance/sonic v1.11.6/go.mod h1:LysEHSvpvDySVdC2f87zGWf6CIKJcAvqab1ZaiQtds4=
+github.com/bytedance/sonic/loader v0.1.1 h1:c+e5Pt1k/cy5wMveRDyk2X4B9hF4g7an8N3zCYjJFNM=
+github.com/bytedance/sonic/loader v0.1.1/go.mod h1:ncP89zfokxS5LZrJxl5z0UJcsk4M4yY2JpfqGeCtNLU=
+github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
+github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
+github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
+github.com/cloudwego/base64x v0.1.4 h1:jwCgWpFanWmN8xoIUHa2rtzmkd5J2plF/dnLS6Xd/0Y=
+github.com/cloudwego/base64x v0.1.4/go.mod h1:0zlkT4Wn5C6NdauXdJRhSKRlJvmclQ1hhJgA0rcu/8w=
+github.com/cloudwego/iasm v0.2.0 h1:1KNIy1I1H9hNNFEEH3DVnI4UujN+1zjpuk6gwHLTssg=
+github.com/cloudwego/iasm v0.2.0/go.mod h1:8rXZaNYT2n95jn+zTI1sDr+IgcD2GVs0nlbbQPiEFhY=
+github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
 github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/rVNCu3HqELle0jiPLLBs70cWOduZpkS1E78=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
-github.com/dlclark/regexp2 v1.10.0 h1:+/GIL799phkJqYW+3YbOd8LCcbHzT0Pbo8zl70MHsq0=
-github.com/dlclark/regexp2 v1.10.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
-github.com/fsnotify/fsnotify v1.4.9 h1:hsms1Qyu0jgnwNXIxa+/V/PDsU6CfLf6CNO8H7IWoS4=
-github.com/gabriel-vasile/mimetype v1.4.2 h1:w5qFW6JKBz9Y393Y4q372O9A7cUSequkh1Q7OhCmWKU=
-github.com/gabriel-vasile/mimetype v1.4.2/go.mod h1:zApsH/mKG4w07erKIaJPFiX0Tsq9BFQgN3qGY5GnNgA=
-github.com/gin-contrib/cors v1.4.0 h1:oJ6gwtUl3lqV0WEIwM/LxPF1QZ5qe2lGWdY2+bz7y0g=
-github.com/gin-contrib/cors v1.4.0/go.mod h1:bs9pNM0x/UsmHPBWT2xZz9ROh8xYjYkiURUfmBoMlcs=
-github.com/gin-contrib/gzip v0.0.6 h1:NjcunTcGAj5CO1gn4N8jHOSIeRFHIbn51z6K+xaN4d4=
-github.com/gin-contrib/gzip v0.0.6/go.mod h1:QOJlmV2xmayAjkNS2Y8NQsMneuRShOU/kjovCXNuzzk=
-github.com/gin-contrib/sessions v0.0.5 h1:CATtfHmLMQrMNpJRgzjWXD7worTh7g7ritsQfmF+0jE=
-github.com/gin-contrib/sessions v0.0.5/go.mod h1:vYAuaUPqie3WUSsft6HUlCjlwwoJQs97miaG2+7neKY=
+github.com/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
+github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
+github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
+github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
+github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
+github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
+github.com/fsnotify/fsnotify v1.7.0 h1:8JEhPFa5W2WU7YfeZzPNqzMP6Lwt7L2715Ggo0nosvA=
+github.com/fsnotify/fsnotify v1.7.0/go.mod h1:40Bi/Hjc2AVfZrqy+aj+yEI+/bRxZnMJyTJwOpGvigM=
+github.com/gabriel-vasile/mimetype v1.4.3 h1:in2uUcidCuFcDKtdcBxlR0rJ1+fsokWf+uqxgUFjbI0=
+github.com/gabriel-vasile/mimetype v1.4.3/go.mod h1:d8uq/6HKRL6CGdk+aubisF/M5GcPfT7nKyLpA0lbSSk=
+github.com/gin-contrib/cors v1.7.2 h1:oLDHxdg8W/XDoN/8zamqk/Drgt4oVZDvaV0YmvVICQw=
+github.com/gin-contrib/cors v1.7.2/go.mod h1:SUJVARKgQ40dmrzgXEVxj2m7Ig1v1qIboQkPDTQ9t2E=
+github.com/gin-contrib/gzip v1.0.1 h1:HQ8ENHODeLY7a4g1Au/46Z92bdGFl74OhxcZble9WJE=
+github.com/gin-contrib/gzip v1.0.1/go.mod h1:njt428fdUNRvjuJf16tZMYZ2Yl+WQB53X5wmhDwXvC4=
+github.com/gin-contrib/sessions v1.0.1 h1:3hsJyNs7v7N8OtelFmYXFrulAf6zSR7nW/putcPEHxI=
+github.com/gin-contrib/sessions v1.0.1/go.mod h1:ouxSFM24/OgIud5MJYQJLpy6AwxQ5EYO9yLhbtObGkM=
 github.com/gin-contrib/sse v0.1.0 h1:Y/yl/+YNO8GZSjAhjMsSuLt29uWRFHdHYUb5lYOV9qE=
 github.com/gin-contrib/sse v0.1.0/go.mod h1:RHrZQHXnP2xjPF+u1gW/2HnVO7nvIa9PG3Gm+fLHvGI=
-github.com/gin-contrib/static v0.0.1 h1:JVxuvHPuUfkoul12N7dtQw7KRn/pSMq7Ue1Va9Swm1U=
-github.com/gin-contrib/static v0.0.1/go.mod h1:CSxeF+wep05e0kCOsqWdAWbSszmc31zTIbD8TvWl7Hs=
-github.com/gin-gonic/gin v1.6.3/go.mod h1:75u5sXoLsGZoRN5Sgbi1eraJ4GU3++wFwWzhwvtwp4M=
-github.com/gin-gonic/gin v1.8.1/go.mod h1:ji8BvRH1azfM+SYow9zQ6SZMvR8qOMZHmsCuWR9tTTk=
-github.com/gin-gonic/gin v1.9.1 h1:4idEAncQnU5cB7BeOkPtxjfCSye0AAm1R0RVIqJ+Jmg=
-github.com/gin-gonic/gin v1.9.1/go.mod h1:hPrL7YrpYKXt5YId3A/Tnip5kqbEAP+KLuI3SUcPTeU=
-github.com/go-playground/assert/v2 v2.0.1/go.mod h1:VDjEfimB/XKnb+ZQfWdccd7VUvScMdVu0Titje2rxJ4=
+github.com/gin-contrib/static v1.1.2 h1:c3kT4bFkUJn2aoRU3s6XnMjJT8J6nNWJkR0NglqmlZ4=
+github.com/gin-contrib/static v1.1.2/go.mod h1:Fw90ozjHCmZBWbgrsqrDvO28YbhKEKzKp8GixhR4yLw=
+github.com/gin-gonic/gin v1.10.0 h1:nTuyha1TYqgedzytsKYqna+DfLos46nTv2ygFy86HFU=
+github.com/gin-gonic/gin v1.10.0/go.mod h1:4PMNQiOhvDRa013RKVbsiNwoyezlm2rm0uX/T7kzp5Y=
+github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
+github.com/go-logr/logr v1.4.1 h1:pKouT5E8xu9zeFC39JXRDukb6JFQPXM5p5I91188VAQ=
+github.com/go-logr/logr v1.4.1/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
+github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
+github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
 github.com/go-playground/assert/v2 v2.2.0 h1:JvknZsQTYeFEAhQwI4qEt9cyV5ONwRHC+lYKSsYSR8s=
-github.com/go-playground/locales v0.13.0/go.mod h1:taPMhCMXrRLJO55olJkUXHZBHCxTMfnGwq/HNwmWNS8=
-github.com/go-playground/locales v0.14.0/go.mod h1:sawfccIbzZTqEDETgFXqTho0QybSa7l++s0DH+LDiLs=
 github.com/go-playground/locales v0.14.1 h1:EWaQ/wswjilfKLTECiXz7Rh+3BjFhfDFKv/oXslEjJA=
 github.com/go-playground/locales v0.14.1/go.mod h1:hxrqLVvrK65+Rwrd5Fc6F2O76J/NuW9t0sjnWqG1slY=
-github.com/go-playground/universal-translator v0.17.0/go.mod h1:UkSxE5sNxxRwHyU+Scu5vgOQjsIJAF8j9muTVoKLVtA=
-github.com/go-playground/universal-translator v0.18.0/go.mod h1:UvRDBj+xPUEGrFYl+lu/H90nyDXpg0fqeB/AQUGNTVA=
 github.com/go-playground/universal-translator v0.18.1 h1:Bcnm0ZwsGyWbCzImXv+pAJnYK9S473LQFuzCbDbfSFY=
 github.com/go-playground/universal-translator v0.18.1/go.mod h1:xekY+UJKNuX9WP91TpwSH2VMlDf28Uj24BCp08ZFTUY=
-github.com/go-playground/validator/v10 v10.2.0/go.mod h1:uOYAAleCW8F/7oMFd6aG0GOhaH6EGOAJShg8Id5JGkI=
-github.com/go-playground/validator/v10 v10.10.0/go.mod h1:74x4gJWsvQexRdW8Pn3dXSGrTK4nAUsbPlLADvpJkos=
-github.com/go-playground/validator/v10 v10.14.0 h1:vgvQWe3XCz3gIeFDm/HnTIbj6UGmg/+t63MyGU2n5js=
-github.com/go-playground/validator/v10 v10.14.0/go.mod h1:9iXMNT7sEkjXb0I+enO7QXmzG6QCsPWY4zveKFVRSyU=
+github.com/go-playground/validator/v10 v10.20.0 h1:K9ISHbSaI0lyB2eWMPJo+kOS/FBExVwjEviJTixqxL8=
+github.com/go-playground/validator/v10 v10.20.0/go.mod h1:dbuPbCMFw/DrkbEynArYaCwl3amGuJotoKCe95atGMM=
 github.com/go-redis/redis/v8 v8.11.5 h1:AcZZR7igkdvfVmQTPnu9WE37LRrO/YrBH5zWyjDC0oI=
 github.com/go-redis/redis/v8 v8.11.5/go.mod h1:gREzHqY1hg6oD9ngVRbLStwAWKhA0FEgq8Jd4h5lpwo=
-github.com/go-sql-driver/mysql v1.6.0 h1:BCTh4TKNUYmOmMUcQ3IipzF5prigylS7XXjEkfCHuOE=
-github.com/go-sql-driver/mysql v1.6.0/go.mod h1:DCzpHaOWr8IXmIStZouvnhqoel9Qv2LBy8hT2VhHyBg=
-github.com/goccy/go-json v0.9.7/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MGFi0w8I=
-github.com/goccy/go-json v0.10.2 h1:CrxCmQqYDkv1z7lO7Wbh2HN93uovUHgrECaO5ZrCXAU=
-github.com/goccy/go-json v0.10.2/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MGFi0w8I=
+github.com/go-sql-driver/mysql v1.7.0/go.mod h1:OXbVy3sEdcQ2Doequ6Z5BW6fXNQTmx+9S1MCJN5yJMI=
+github.com/go-sql-driver/mysql v1.8.1 h1:LedoTUt/eveggdHS9qUFC1EFSa8bU2+1pZjSRpvNJ1Y=
+github.com/go-sql-driver/mysql v1.8.1/go.mod h1:wEBSXgmK//2ZFJyE+qWnIsVGmvmEKlqwuVSjsCm7DZg=
+github.com/goccy/go-json v0.10.3 h1:KZ5WoDbxAIgm2HNbYckL0se1fHD6rz5j4ywS6ebzDqA=
+github.com/goccy/go-json v0.10.3/go.mod h1:oq7eo15ShAhp70Anwd5lgX2pLfOS3QCiwU/PULtXL6M=
 github.com/golang-jwt/jwt v3.2.2+incompatible h1:IfV12K8xAKAnZqdXVzCZ+TOjboZ2keLg81eXfW3O+oY=
 github.com/golang-jwt/jwt v3.2.2+incompatible/go.mod h1:8pz2t5EyA70fFQQSrl6XZXzqecmYZeUEB8OUGHkxJ+I=
-github.com/golang/protobuf v1.3.3/go.mod h1:vzj43D7+SQXF/4pzW/hwtAqwc6iTitCiVSaWz5lYuqw=
-github.com/golang/protobuf v1.5.0/go.mod h1:FsONVRAS9T7sI+LIUmWTfcYkHO4aIWwzhcaSAoJOfIk=
-github.com/google/go-cmp v0.5.5 h1:Khx7svrCpmxxtHBq5j2mp/xVjsi8hQMfNLvJFAlrGgU=
-github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
+github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da h1:oI5xCqsCo564l8iNU+DwB5epxmsaqB+rhGL0m5jtYqE=
+github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
+github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
+github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
+github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
+github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
+github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
+github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
+github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
+github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
+github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
+github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
+github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
+github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
+github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.5.3/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/go-cmp v0.6.0 h1:ofyhxvXcZhMsU5ulbFiLKl/XBFqE1GSq7atu8tAmTRI=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
-github.com/google/uuid v1.3.0 h1:t6JiXgmwXMjEs8VusXIJk2BXHsn+wx8BZdTaoZ5fu7I=
-github.com/google/uuid v1.3.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
-github.com/gorilla/context v1.1.1 h1:AWwleXJkX/nhcU9bZSnZoi3h/qGYqQAGhq6zZe/aQW8=
-github.com/gorilla/context v1.1.1/go.mod h1:kBGZzfjB9CEq2AlWe17Uuf7NDRt0dE0s8S51q0aT7Yg=
-github.com/gorilla/securecookie v1.1.1 h1:miw7JPhV+b/lAHSXz4qd/nN9jRiAFV5FwjeKyCS8BvQ=
-github.com/gorilla/securecookie v1.1.1/go.mod h1:ra0sb63/xPlUeL+yeDciTfxMRAA+MP+HVt/4epWDjd4=
-github.com/gorilla/sessions v1.2.1 h1:DHd3rPN5lE3Ts3D8rKkQ8x/0kqfeNmBAaiSi+o7FsgI=
-github.com/gorilla/sessions v1.2.1/go.mod h1:dk2InVEVJ0sfLlnXv9EAgkf6ecYs/i80K/zI+bUmuGM=
-github.com/gorilla/websocket v1.5.0 h1:PPwGk2jz7EePpoHN/+ClbZu8SPxiqlu12wZP/3sWmnc=
-github.com/gorilla/websocket v1.5.0/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
+github.com/google/gofuzz v1.2.0 h1:xRy4A+RhZaiKjJ1bPfwQ8sedCA+YS2YcCHW6ec7JMi0=
+github.com/google/s2a-go v0.1.7 h1:60BLSyTrOV4/haCDW4zb1guZItoSq8foHCXrAnjBo/o=
+github.com/google/s2a-go v0.1.7/go.mod h1:50CgR4k1jNlWBu4UfS4AcfhVe1r6pdZPygJ3R8F0Qdw=
+github.com/google/uuid v1.1.2/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
+github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2 h1:Vie5ybvEvT75RniqhfFxPRy3Bf7vr3h0cechB90XaQs=
+github.com/googleapis/enterprise-certificate-proxy v0.3.2/go.mod h1:VLSiSSBs/ksPL8kq3OBOQ6WRI2QnaFynd1DCjZ62+V0=
+github.com/googleapis/gax-go/v2 v2.12.5 h1:8gw9KZK8TiVKB6q3zHY3SBzLnrGp6HQjyfYBYGmXdxA=
+github.com/googleapis/gax-go/v2 v2.12.5/go.mod h1:BUDKcWo+RaKq5SC9vVYL0wLADa3VcfswbOMMRmB9H3E=
+github.com/gopherjs/gopherjs v1.17.2 h1:fQnZVsXk8uxXIStYb0N4bGk7jeyTalG/wsZjQ25dO0g=
+github.com/gopherjs/gopherjs v1.17.2/go.mod h1:pRRIvn/QzFLrKfvEz3qUuEhtE/zLCWfreZ6J5gM2i+k=
+github.com/gorilla/context v1.1.2 h1:WRkNAv2uoa03QNIc1A6u4O7DAGMUVoopZhkiXWA2V1o=
+github.com/gorilla/context v1.1.2/go.mod h1:KDPwT9i/MeWHiLl90fuTgrt4/wPcv75vFAZLaOOcbxM=
+github.com/gorilla/securecookie v1.1.2 h1:YCIWL56dvtr73r6715mJs5ZvhtnY73hBvEF8kXD8ePA=
+github.com/gorilla/securecookie v1.1.2/go.mod h1:NfCASbcHqRSY+3a8tlWJwsQap2VX5pwzwo4h3eOamfo=
+github.com/gorilla/sessions v1.2.2 h1:lqzMYz6bOfvn2WriPUjNByzeXIlVzURcPmgMczkmTjY=
+github.com/gorilla/sessions v1.2.2/go.mod h1:ePLdVu+jbEgHH+KWw8I1z2wqd0BAdAQh/8LRvBeoNcQ=
+github.com/gorilla/websocket v1.5.1 h1:gmztn0JnHVt9JZquRuzLw3g4wouNVzKL15iLr/zn/QY=
+github.com/gorilla/websocket v1.5.1/go.mod h1:x3kM2JMyaluk02fnUJpQuwD2dCS5NDG2ZHL0uE0tcaY=
 github.com/jackc/pgpassfile v1.0.0 h1:/6Hmqy13Ss2zCq62VdNG8tM1wchn8zjSGOBJ6icpsIM=
 github.com/jackc/pgpassfile v1.0.0/go.mod h1:CEx0iS5ambNFdcRtxPj5JhEz+xB6uRky5eyVu/W2HEg=
-github.com/jackc/pgservicefile v0.0.0-20221227161230-091c0ba34f0a h1:bbPeKD0xmW/Y25WS6cokEszi5g+S0QxI/d45PkRi7Nk=
-github.com/jackc/pgservicefile v0.0.0-20221227161230-091c0ba34f0a/go.mod h1:5TJZWKEWniPve33vlWYSoGYefn3gLQRzjfDlhSJ9ZKM=
-github.com/jackc/pgx/v5 v5.5.4 h1:Xp2aQS8uXButQdnCMWNmvx6UysWQQC+u1EoizjguY+8=
-github.com/jackc/pgx/v5 v5.5.4/go.mod h1:ez9gk+OAat140fv9ErkZDYFWmXLfV+++K0uAOiwgm1A=
+github.com/jackc/pgservicefile v0.0.0-20231201235250-de7065d80cb9 h1:L0QtFUgDarD7Fpv9jeVMgy/+Ec0mtnmYuImjTz6dtDA=
+github.com/jackc/pgservicefile v0.0.0-20231201235250-de7065d80cb9/go.mod h1:5TJZWKEWniPve33vlWYSoGYefn3gLQRzjfDlhSJ9ZKM=
+github.com/jackc/pgx/v5 v5.5.5 h1:amBjrZVmksIdNjxGW/IiIMzxMKZFelXbUoPNb+8sjQw=
+github.com/jackc/pgx/v5 v5.5.5/go.mod h1:ez9gk+OAat140fv9ErkZDYFWmXLfV+++K0uAOiwgm1A=
 github.com/jackc/puddle/v2 v2.2.1 h1:RhxXJtFG022u4ibrCSMSiu5aOq1i77R3OHKNJj77OAk=
 github.com/jackc/puddle/v2 v2.2.1/go.mod h1:vriiEXHvEE654aYKXXjOvZM39qJ0q+azkZFrfEOc3H4=
+github.com/jinzhu/copier v0.4.0 h1:w3ciUoD19shMCRargcpm0cm91ytaBhDvuRpz1ODO/U8=
+github.com/jinzhu/copier v0.4.0/go.mod h1:DfbEm0FYsaqBcKcFuvmOZb218JkPGtvSHsKg8S8hyyg=
 github.com/jinzhu/inflection v1.0.0 h1:K317FqzuhWc8YvSVlFMCCUb36O/S9MCKRDI7QkRKD/E=
 github.com/jinzhu/inflection v1.0.0/go.mod h1:h+uFLlag+Qp1Va5pdKtLDYj+kHp5pxUVkryuEj+Srlc=
-github.com/jinzhu/now v1.1.4/go.mod h1:d3SSVoowX0Lcu0IBviAWJpolVfI5UJVZZ7cO71lE/z8=
 github.com/jinzhu/now v1.1.5 h1:/o9tlHleP7gOFmsnYNz3RGnqzefHA47wQpKrrdTIwXQ=
 github.com/jinzhu/now v1.1.5/go.mod h1:d3SSVoowX0Lcu0IBviAWJpolVfI5UJVZZ7cO71lE/z8=
-github.com/json-iterator/go v1.1.9/go.mod h1:KdQUCv79m/52Kvf8AW2vK1V8akMuk1QjK/uOdHXbAo4=
+github.com/joho/godotenv v1.5.1 h1:7eLL/+HRGLY0ldzfGMeQkb7vMd0as4CfYvUVzLqw0N0=
+github.com/joho/godotenv v1.5.1/go.mod h1:f4LDr5Voq0i2e/R5DDNOoa2zzDfwtkZa6DnEwAbqwq4=
 github.com/json-iterator/go v1.1.12 h1:PV8peI4a0ysnczrg+LtxykD8LfKY9ML6u2jnxaEnrnM=
 github.com/json-iterator/go v1.1.12/go.mod h1:e30LSqwooZae/UwlEbR2852Gd8hjQvJoHmT4TnhNGBo=
+github.com/jtolds/gls v4.20.0+incompatible h1:xdiiI2gbIgH/gLH7ADydsJ1uDOEzR8yvV7C0MuV77Wo=
+github.com/jtolds/gls v4.20.0+incompatible/go.mod h1:QJZ7F/aHp+rZTRtaJ1ow/lLfFfVYBRgL+9YlvaHOwJU=
 github.com/klauspost/cpuid/v2 v2.0.9/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg=
-github.com/klauspost/cpuid/v2 v2.2.4 h1:acbojRNwl3o09bUq+yDCtZFc1aiwaAAxtcn8YkZXnvk=
-github.com/klauspost/cpuid/v2 v2.2.4/go.mod h1:RVVoqg1df56z8g3pUjL/3lE5UfnlrJX8tyFgg4nqhuY=
-github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
-github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
+github.com/klauspost/cpuid/v2 v2.2.7 h1:ZWSB3igEs+d0qvnxR/ZBzXVmxkgt8DdzP6m9pfuVLDM=
+github.com/klauspost/cpuid/v2 v2.2.7/go.mod h1:Lcz8mBdAVJIBVzewtcLocK12l3Y+JytZYpaMropDUws=
+github.com/knz/go-libedit v1.10.1/go.mod h1:MZTVkCWyz0oBc7JOWP3wNAzd002ZbM/5hgShxwh4x8M=
 github.com/kr/pretty v0.3.0 h1:WgNl7dwNpEZ6jJ9k1snq4pZsg7DOEN8hP9Xw0Tsjwk0=
-github.com/kr/pretty v0.3.0/go.mod h1:640gp4NfQd8pI5XOwp5fnNeVWj67G7CFk/SaSQn7NBk=
-github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
-github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
 github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
 github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
-github.com/leodido/go-urn v1.2.0/go.mod h1:+8+nEpDfqqsY+g338gtMEUOtuK+4dEMhiQEgxpxOKII=
-github.com/leodido/go-urn v1.2.1/go.mod h1:zt4jvISO2HfUBqxjfIshjdMTYS56ZS/qv49ictyFfxY=
-github.com/leodido/go-urn v1.2.4 h1:XlAE/cm/ms7TE/VMVoduSpNBoyc2dOxHs5MZSwAN63Q=
-github.com/leodido/go-urn v1.2.4/go.mod h1:7ZrI8mTSeBSHl/UaRyKQW1qZeMgak41ANeCNaVckg+4=
-github.com/mattn/go-isatty v0.0.12/go.mod h1:cbi8OIDigv2wuxKPP5vlRcQ1OAZbq2CE4Kysco4FUpU=
-github.com/mattn/go-isatty v0.0.14/go.mod h1:7GGIvUiUoEMVVmxf/4nioHXj79iQHKdU27kJ6hsGG94=
-github.com/mattn/go-isatty v0.0.19 h1:JITubQf0MOLdlGRuRq+jtsDlekdYPia9ZFsB8h/APPA=
-github.com/mattn/go-isatty v0.0.19/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
-github.com/mattn/go-sqlite3 v1.14.15/go.mod h1:2eHXhiwb8IkHr+BDWZGa96P6+rkvnG63S2DGjv9HUNg=
-github.com/mattn/go-sqlite3 v2.0.3+incompatible h1:gXHsfypPkaMZrKbD5209QV9jbUTJKjyR5WD3HYQSd+U=
-github.com/mattn/go-sqlite3 v2.0.3+incompatible/go.mod h1:FPy6KqzDD04eiIsT53CuJW3U88zkxoIYsOqkbpncsNc=
+github.com/leodido/go-urn v1.4.0 h1:WT9HwE9SGECu3lg4d/dIA+jxlljEa1/ffXKmRjqdmIQ=
+github.com/leodido/go-urn v1.4.0/go.mod h1:bvxc+MVxLKB4z00jd1z+Dvzr47oO32F/QSNjSBOlFxI=
+github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
+github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
+github.com/mattn/go-sqlite3 v1.14.22 h1:2gZY6PC6kBnID23Tichd1K+Z0oS6nE/XwU+Vz/5o4kU=
+github.com/mattn/go-sqlite3 v1.14.22/go.mod h1:Uh1q+B4BYcTPb+yiD3kU8Ct7aC0hY9fxUwlHK0RXw+Y=
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
-github.com/modern-go/reflect2 v0.0.0-20180701023420-4b7aa43c6742/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
 github.com/modern-go/reflect2 v1.0.2 h1:xBagoLtFs94CBntxluKeaWgTMpvLxC4ur3nMaC9Gz0M=
 github.com/modern-go/reflect2 v1.0.2/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjYzDa0/r8luk=
 github.com/nxadm/tail v1.4.8 h1:nPr65rt6Y5JFSKQO7qToXr7pePgD6Gwiw05lkbyAQTE=
 github.com/onsi/ginkgo v1.16.5 h1:8xi0RTUf59SOSfEtZMvwTvXYMzG4gV23XVHOZiXNtnE=
 github.com/onsi/gomega v1.18.1 h1:M1GfJqGRrBrrGGsbxzV5dqM2U2ApXefZCQpkukxYRLE=
-github.com/pelletier/go-toml/v2 v2.0.1/go.mod h1:r9LEWfGN8R5k0VXJ+0BkIe7MYkRdwZOjgMj2KwnJFUo=
-github.com/pelletier/go-toml/v2 v2.0.8 h1:0ctb6s9mE31h0/lhu+J6OPmVeDxJn+kYnJc2jZR9tGQ=
-github.com/pelletier/go-toml/v2 v2.0.8/go.mod h1:vuYfssBdrU2XDZ9bYydBu6t+6a6PYNcZljzZR9VXg+4=
-github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsKMEsNbx1VGcRFpLqf3715MtcvvzbA=
-github.com/pkoukk/tiktoken-go v0.1.5 h1:hAlT4dCf6Uk50x8E7HQrddhH3EWMKUN+LArExQQsQx4=
-github.com/pkoukk/tiktoken-go v0.1.5/go.mod h1:9NiV+i9mJKGj1rYOT+njbv+ZwA/zJxYdewGl6qVatpg=
+github.com/patrickmn/go-cache v2.1.0+incompatible h1:HRMgzkcYKYpi3C8ajMPV8OFXaaRUnok+kx1WdO15EQc=
+github.com/patrickmn/go-cache v2.1.0+incompatible/go.mod h1:3Qf8kWWT7OJRJbdiICTKqZju1ZixQ/KpMGzzAfe6+WQ=
+github.com/pelletier/go-toml/v2 v2.2.2 h1:aYUidT7k73Pcl9nb2gScu7NSrKCSHIDE89b3+6Wq+LM=
+github.com/pelletier/go-toml/v2 v2.2.2/go.mod h1:1t835xjRzz80PqgE6HHgN2JOsmgYu/h4qDAS4n929Rs=
+github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
+github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
+github.com/pkoukk/tiktoken-go v0.1.7 h1:qOBHXX4PHtvIvmOtyg1EeKlwFRiMKAcoMp4Q+bLQDmw=
+github.com/pkoukk/tiktoken-go v0.1.7/go.mod h1:9NiV+i9mJKGj1rYOT+njbv+ZwA/zJxYdewGl6qVatpg=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/rogpeppe/go-internal v1.6.1/go.mod h1:xXDCJY+GAPziupqXw64V24skbSoqbTEfhy4qGm1nDQc=
+github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
 github.com/rogpeppe/go-internal v1.8.0 h1:FCbCCtXNOY3UtUuHUYaghJg4y7Fd14rXifAYUAtL9R8=
-github.com/rogpeppe/go-internal v1.8.0/go.mod h1:WmiCO8CzOY8rg0OYDC4/i/2WRWAB6poM+XZ2dLUbcbE=
+github.com/smarty/assertions v1.15.0 h1:cR//PqUBUiQRakZWqBiFFQ9wb8emQGDb0HeGdqGByCY=
+github.com/smarty/assertions v1.15.0/go.mod h1:yABtdzeQs6l1brC900WlRNwj6ZR55d7B+E8C6HtKdec=
+github.com/smartystreets/goconvey v1.8.1 h1:qGjIddxOk4grTu9JPOU31tVfq3cNdBlNa5sSznIX1xY=
+github.com/smartystreets/goconvey v1.8.1/go.mod h1:+/u4qLyY6x1jReYOp7GOM2FSt8aP9CzCZL03bI28W60=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/objx v0.4.0/go.mod h1:YvHI0jy2hoMjB+UWwv71VJQ9isScKT/TqJzVSSt89Yw=
 github.com/stretchr/objx v0.5.0/go.mod h1:Yh+to48EsGEfYuaHDzXPcE3xhTkx73EhmCGUpEOglKo=
+github.com/stretchr/objx v0.5.2/go.mod h1:FRsXN1f5AsAjCGJKqEizvkpNtU+EGNCLh3NxZ/8L+MA=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
-github.com/stretchr/testify v1.4.0/go.mod h1:j7eGeouHqKxXV5pUuKE4zz7dFj8WfuZ+81PSLYec5m4=
-github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
 github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
 github.com/stretchr/testify v1.7.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
 github.com/stretchr/testify v1.8.0/go.mod h1:yNjHg4UonilssWZ8iaSj1OCr/vHnekPRkoO+kdMU+MU=
 github.com/stretchr/testify v1.8.1/go.mod h1:w2LPCIKwWwSfY2zedu0+kehJoqGctiVI29o6fzry7u4=
-github.com/stretchr/testify v1.8.2/go.mod h1:w2LPCIKwWwSfY2zedu0+kehJoqGctiVI29o6fzry7u4=
-github.com/stretchr/testify v1.8.3 h1:RP3t2pwF7cMEbC1dqtB6poj3niw/9gnV4Cjg5oW5gtY=
-github.com/stretchr/testify v1.8.3/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
+github.com/stretchr/testify v1.8.4/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
+github.com/stretchr/testify v1.9.0 h1:HtqpIVDClZ4nwg75+f6Lvsy/wHu+3BoSGCbBAcpTsTg=
+github.com/stretchr/testify v1.9.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
 github.com/twitchyliquid64/golang-asm v0.15.1 h1:SU5vSMR7hnwNxj24w34ZyCi/FmDZTkS4MhqMhdFk5YI=
 github.com/twitchyliquid64/golang-asm v0.15.1/go.mod h1:a1lVb/DtPvCB8fslRZhAngC2+aY1QWCk3Cedj/Gdt08=
-github.com/ugorji/go v1.1.7/go.mod h1:kZn38zHttfInRq0xu/PH0az30d+z6vm202qpg1oXVMw=
-github.com/ugorji/go v1.2.7/go.mod h1:nF9osbDWLy6bDVv/Rtoh6QgnvNDpmCalQV5urGCCS6M=
-github.com/ugorji/go/codec v1.1.7/go.mod h1:Ax+UKWsSmolVDwsd+7N3ZtXu+yMGCf907BLYF3GoBXY=
-github.com/ugorji/go/codec v1.2.7/go.mod h1:WGN1fab3R1fzQlVQTkfxVtIBhWDRqOviHU95kRgeqEY=
-github.com/ugorji/go/codec v1.2.11 h1:BMaWp1Bb6fHwEtbplGBGJ498wD+LKlNSl25MjdZY4dU=
-github.com/ugorji/go/codec v1.2.11/go.mod h1:UNopzCgEMSXjBc6AOMqYvWC1ktqTAfzJZUZgYf6w6lg=
+github.com/ugorji/go/codec v1.2.12 h1:9LC83zGrHhuUA9l16C9AHXAqEV/2wBQ4nkvumAE65EE=
+github.com/ugorji/go/codec v1.2.12/go.mod h1:UNopzCgEMSXjBc6AOMqYvWC1ktqTAfzJZUZgYf6w6lg=
+go.opencensus.io v0.24.0 h1:y73uSU6J157QMP2kn2r30vwW1A2W2WFwSCGnAVxeaD0=
+go.opencensus.io v0.24.0/go.mod h1:vNK8G9p7aAivkbmorf4v+7Hgx+Zs0yY+0fOtgBfjQKo=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0 h1:4Pp6oUg3+e/6M4C0A/3kJ2VYa++dsWVTtGgLVj5xtHg=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.49.0/go.mod h1:Mjt1i1INqiaoZOMGR1RIUJN+i3ChKoFRqzrRQhlkbs0=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 h1:jq9TW8u3so/bN+JPT166wjOI6/vQPF6Xe7nMNIltagk=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0/go.mod h1:p8pYQP+m5XfbZm9fxtSKAbM6oIllS7s2AfxrChvc7iw=
+go.opentelemetry.io/otel v1.24.0 h1:0LAOdjNmQeSTzGBzduGe/rU4tZhMwL5rWgtp9Ku5Jfo=
+go.opentelemetry.io/otel v1.24.0/go.mod h1:W7b9Ozg4nkF5tWI5zsXkaKKDjdVjpD4oAt9Qi/MArHo=
+go.opentelemetry.io/otel/metric v1.24.0 h1:6EhoGWWK28x1fbpA4tYTOWBkPefTDQnb8WSGXlc88kI=
+go.opentelemetry.io/otel/metric v1.24.0/go.mod h1:VYhLe1rFfxuTXLgj4CBiyz+9WYBA8pNGJgDcSFRKBco=
+go.opentelemetry.io/otel/trace v1.24.0 h1:CsKnnL4dUAr/0llH9FKuc698G04IrpWV0MQA/Y1YELI=
+go.opentelemetry.io/otel/trace v1.24.0/go.mod h1:HPc3Xr/cOApsBI154IU0OI0HJexz+aw5uPdbs3UCjNU=
 golang.org/x/arch v0.0.0-20210923205945-b76863e36670/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
-golang.org/x/arch v0.3.0 h1:02VY4/ZcO/gBOH6PUaoiptASxtXU10jazRCP865E97k=
-golang.org/x/arch v0.3.0/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
-golang.org/x/crypto v0.0.0-20210711020723-a769d52b0f97/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
-golang.org/x/crypto v0.17.0 h1:r8bRNjWL3GshPW3gkd+RpvzWrZAwPS49OmTGZ/uhM4k=
-golang.org/x/crypto v0.17.0/go.mod h1:gCAAfMLgwOJRpTjQ2zCCt2OcSfYMTeZVSRtQlPC7Nq4=
-golang.org/x/image v0.14.0 h1:tNgSxAFe3jC4uYqvZdTr84SZoM1KfwdC9SKIFrLjFn4=
-golang.org/x/image v0.14.0/go.mod h1:HUYqC05R2ZcZ3ejNQsIHQDQiwWM4JBqmm6MKANTp4LE=
-golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.17.0 h1:pVaXccu2ozPjCXewfr1S7xza/zcXTity9cCdXQYSjIM=
-golang.org/x/net v0.17.0/go.mod h1:NxSsAGuq816PNPmqtQdLE42eU2Fs7NoRIZrHJAlaCOE=
-golang.org/x/sync v0.1.0 h1:wsuoTGHzEhffawBOhz5CYhcrV4IdKZbEyZjBMuTp12o=
-golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sys v0.0.0-20200116001909-b77594299b42/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210630005230-0f9fa26af87c/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210806184541-e5e7981a1069/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20220704084225-05e143d24a9e/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/arch v0.8.0 h1:3wRIsP3pM4yUptoR96otTUOXI367OS0+c9eeRi9doIc=
+golang.org/x/arch v0.8.0/go.mod h1:FEVrYAQjsQXMVJ1nsMoVVXPZg6p2JE2mx8psSWTDQys=
+golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
+golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
+golang.org/x/crypto v0.31.0 h1:ihbySMvVjLAeSH1IbfcRTkD/iNscyz8rGzjF/E5hV6U=
+golang.org/x/crypto v0.31.0/go.mod h1:kDsLvtWBEx7MV9tJOj9bnXsPbxwJQ6csT/x4KIN4Ssk=
+golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
+golang.org/x/image v0.18.0 h1:jGzIakQa/ZXI1I0Fxvaa9W7yP25TqT6cHIHn+6CqvSQ=
+golang.org/x/image v0.18.0/go.mod h1:4yyo5vMFQjVjUcVk4jEQcU9MGy/rulF5WvUILseCM2E=
+golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
+golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
+golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
+golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
+golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
+golang.org/x/net v0.26.0 h1:soB7SVo0PWrY4vPW/+ay0jKDNScG2X9wFeYlXIvJsOQ=
+golang.org/x/net v0.26.0/go.mod h1:5YKkiSynbBIh3p6iOc/vibscux0x38BZDkn8sCUPxHE=
+golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
+golang.org/x/oauth2 v0.21.0 h1:tsimM75w1tF/uws5rbeHzIWxEqElMehnc+iW793zsZs=
+golang.org/x/oauth2 v0.21.0/go.mod h1:XYTD2NtWslqkgxebSiOHnXEap4TF09sJSc7H1sXbhtI=
+golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.10.0 h1:3NQrjDixjgGwUOCaF8w2+VYHv0Ve/vGYSbdkTa98gmQ=
+golang.org/x/sync v0.10.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
+golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.5.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.15.0 h1:h48lPFYpsTvQJZF4EKyI4aLHaev3CxivZmv7yZig9pc=
-golang.org/x/sys v0.15.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
-golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
+golang.org/x/sys v0.28.0 h1:Fksou7UEQUWlKvIdsqzJmUmCX3cZuD2+P3XyyzwMhlA=
+golang.org/x/sys v0.28.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
 golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.14.0 h1:ScX5w1eTa3QqT8oi6+ziP7dTV1S2+ALU0bI+0zXKWiQ=
-golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
+golang.org/x/text v0.21.0 h1:zyQAAkrwaneQ066sspRyJaG9VNi/YJ1NfzcGB3hZ/qo=
+golang.org/x/text v0.21.0/go.mod h1:4IBbMaMmOPCJ8SecivzSH54+73PCFmPWxNTLm+vZkEQ=
+golang.org/x/time v0.5.0 h1:o7cqy6amK/52YcAKIPlM3a+Fpj35zvRj2TP+e1xFSfk=
+golang.org/x/time v0.5.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543 h1:E7g+9GITq07hpfrRu66IVDexMakfv52eLZ2CXBWiKr4=
+golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
+golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
+golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
+golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
-google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=
-google.golang.org/protobuf v1.28.0/go.mod h1:HV8QOd/L58Z+nl8r43ehVNZIU/HEI6OcFqwMG9pJV4I=
-google.golang.org/protobuf v1.33.0 h1:uNO2rsAINq/JlFpSdYEKIZ0uKD/R9cpdv0T+yoGwGmI=
-google.golang.org/protobuf v1.33.0/go.mod h1:c6P6GXX6sHbq/GpV6MGZEdwhWPcYBgnhAHhKbcUYpos=
+google.golang.org/api v0.187.0 h1:Mxs7VATVC2v7CY+7Xwm4ndkX71hpElcvx0D1Ji/p1eo=
+google.golang.org/api v0.187.0/go.mod h1:KIHlTc4x7N7gKKuVsdmfBXN13yEEWXWFURWY6SBp2gk=
+google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
+google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
+google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
+google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
+google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4 h1:MuYw1wJzT+ZkybKfaOXKp5hJiZDn2iHaXRw0mRYdHSc=
+google.golang.org/genproto/googleapis/api v0.0.0-20240617180043-68d350f18fd4/go.mod h1:px9SlOOZBg1wM1zdnr8jEL4CNGUBZ+ZKYtNPApNQc4c=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d h1:k3zyW3BYYR30e8v3x0bTDdE9vpYFjZHK+HcyqkrppWk=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20240624140628-dc46fd24d27d/go.mod h1:Ue6ibwXGpU+dqIcODieyLOcgj7z8+IcskoNIgZxtrFY=
+google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
+google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
+google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=
+google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
+google.golang.org/grpc v1.33.2/go.mod h1:JMHMWHQWaTccqQQlmk3MJZS+GWXOdAesneDmEnv2fbc=
+google.golang.org/grpc v1.64.1 h1:LKtvyfbX3UGVPFcGqJ9ItpVWW6oN/2XqTxfAnwRRXiA=
+google.golang.org/grpc v1.64.1/go.mod h1:hiQF4LFZelK2WKaP6W0L92zGHtiQdZxk8CrSdvyjeP0=
+google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
+google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
+google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
+google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
+google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
+google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
+google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
+google.golang.org/protobuf v1.34.2 h1:6xV6lTsCfpGD21XK49h7MhtcApnLqkfYgPcdHftf6hg=
+google.golang.org/protobuf v1.34.2/go.mod h1:qYOHts0dSfpeUzUFpOMr/WGzszTmLH+DiWniOlNbLDw=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
-gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
-gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
-gopkg.in/errgo.v2 v2.1.0/go.mod h1:hNsd1EY+bozCKY1Ytp96fpM3vjJbqLJn88ws8XvfDNI=
 gopkg.in/tomb.v1 v1.0.0-20141024135613-dd632973f1e7 h1:uRGJdciOHaEIrze2W8Q3AKkepLTh2hOroT7a+7czfdQ=
-gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
-gopkg.in/yaml.v2 v2.4.0/go.mod h1:RDklbk79AGWmwhnvt/jBztapEOGDOx6ZbXqjP6csGnQ=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
-gopkg.in/yaml.v3 v3.0.0-20210107192922-496545a6307b/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
-gorm.io/driver/mysql v1.4.3 h1:/JhWJhO2v17d8hjApTltKNADm7K7YI2ogkR7avJUL3k=
-gorm.io/driver/mysql v1.4.3/go.mod h1:sSIebwZAVPiT+27jK9HIwvsqOGKx3YMPmrA3mBJR10c=
-gorm.io/driver/postgres v1.5.2 h1:ytTDxxEv+MplXOfFe3Lzm7SjG09fcdb3Z/c056DTBx0=
-gorm.io/driver/postgres v1.5.2/go.mod h1:fmpX0m2I1PKuR7mKZiEluwrP3hbs+ps7JIGMUBpCgl8=
-gorm.io/driver/sqlite v1.4.3 h1:HBBcZSDnWi5BW3B3rwvVTc510KGkBkexlOg0QrmLUuU=
-gorm.io/driver/sqlite v1.4.3/go.mod h1:0Aq3iPO+v9ZKbcdiz8gLWRw5VOPcBOPUQJFLq5e2ecI=
-gorm.io/gorm v1.23.8/go.mod h1:l2lP/RyAtc1ynaTjFksBde/O8v9oOGIApu2/xRitmZk=
-gorm.io/gorm v1.24.0/go.mod h1:DVrVomtaYTbqs7gB/x2uVvqnXzv0nqjB396B8cG4dBA=
-gorm.io/gorm v1.25.0 h1:+KtYtb2roDz14EQe4bla8CbQlmb9dN3VejSai3lprfU=
-gorm.io/gorm v1.25.0/go.mod h1:L4uxeKpfBml98NYqVqwAdmV1a2nBtAec/cf3fpucW/k=
+gorm.io/driver/mysql v1.5.6 h1:Ld4mkIickM+EliaQZQx3uOJDJHtrd70MxAUqWqlx3Y8=
+gorm.io/driver/mysql v1.5.6/go.mod h1:sEtPWMiqiN1N1cMXoXmBbd8C6/l+TESwriotuRRpkDM=
+gorm.io/driver/postgres v1.5.7 h1:8ptbNJTDbEmhdr62uReG5BGkdQyeasu/FZHxI0IMGnM=
+gorm.io/driver/postgres v1.5.7/go.mod h1:3e019WlBaYI5o5LIdNV+LyxCMNtLOQETBXL2h4chKpA=
+gorm.io/driver/sqlite v1.5.5 h1:7MDMtUZhV065SilG62E0MquljeArQZNfJnjd9i9gx3E=
+gorm.io/driver/sqlite v1.5.5/go.mod h1:6NgQ7sQWAIFsPrJJl1lSNSu2TABh0ZZ/zm5fosATavE=
+gorm.io/gorm v1.25.7/go.mod h1:hbnx/Oo0ChWMn1BIhpy1oYozzpM15i4YPuHDmfYtwg8=
+gorm.io/gorm v1.25.10 h1:dQpO+33KalOA+aFYGlK+EfxcI5MbO7EP2yYygwh9h+s=
+gorm.io/gorm v1.25.10/go.mod h1:hbnx/Oo0ChWMn1BIhpy1oYozzpM15i4YPuHDmfYtwg8=
+honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
+nullprogram.com/x/optparse v1.0.0/go.mod h1:KdyPE+Igbe0jQUrVfMqDMeJQIJZEuyV7pjYmp6pbG50=
 rsc.io/pdf v0.1.1/go.mod h1:n8OzWcQ6Sp37PL01nO98y4iUCRdTGarVfzxY20ICaU4=
diff --git a/i18n/en.json b/i18n/en.json
index 54728e2f..b7f1bd3e 100644
--- a/i18n/en.json
+++ b/i18n/en.json
@@ -8,12 +8,12 @@
   "确认删除": "Confirm Delete",
   "确认绑定": "Confirm Binding",
   "您正在删除自己的帐户，将清空所有数据且不可恢复": "You are deleting your account, all data will be cleared and unrecoverable.",
-  "\"通道「%s」（#%d）已被禁用\"": "\"Channel %s (#%d) has been disabled\"",
-  "通道「%s」（#%d）已被禁用，原因：%s": "Channel %s (#%d) has been disabled, reason: %s",
+  "\"渠道「%s」（#%d）已被禁用\"": "\"Channel %s (#%d) has been disabled\"",
+  "渠道「%s」（#%d）已被禁用，原因：%s": "Channel %s (#%d) has been disabled, reason: %s",
   "测试已在运行中": "Test is already running",
   "响应时间 %.2fs 超过阈值 %.2fs": "Response time %.2fs exceeds threshold %.2fs",
-  "通道测试完成": "Channel test completed",
-  "通道测试完成，如果没有收到禁用通知，说明所有通道都正常": "Channel test completed, if you have not received the disable notification, it means that all channels are normal",
+  "渠道测试完成": "Channel test completed",
+  "渠道测试完成，如果没有收到禁用通知，说明所有渠道都正常": "Channel test completed, if you have not received the disable notification, it means that all channels are normal",
   "无法连接至 GitHub 服务器，请稍后重试！": "Unable to connect to GitHub server, please try again later!",
   "返回值非法，用户字段为空，请稍后重试！": "The return value is illegal, the user field is empty, please try again later!",
   "管理员未开启通过 GitHub 登录以及注册": "The administrator did not turn on login and registration via GitHub",
@@ -119,11 +119,11 @@
   " 个月 ": " M ",
   " 年 ": " y ",
   "未测试": "Not tested",
-  "通道 ${name} 测试成功，耗时 ${time.toFixed(2)} 秒。": "Channel ${name} test succeeded, time consumed ${time.toFixed(2)} s.",
-  "已成功开始测试所有通道，请刷新页面查看结果。": "All channels have been successfully tested, please refresh the page to view the results.",
-  "已成功开始测试所有已启用通道，请刷新页面查看结果。": "All enabled channels have been successfully tested, please refresh the page to view the results.",
-  "通道 ${name} 余额更新成功！": "Channel ${name} balance updated successfully!",
-  "已更新完毕所有已启用通道余额！": "The balance of all enabled channels has been updated!",
+  "渠道 ${name} 测试成功，耗时 ${time.toFixed(2)} 秒。": "Channel ${name} test succeeded, time consumed ${time.toFixed(2)} s.",
+  "已成功开始测试所有渠道，请刷新页面查看结果。": "All channels have been successfully tested, please refresh the page to view the results.",
+  "已成功开始测试所有已启用渠道，请刷新页面查看结果。": "All enabled channels have been successfully tested, please refresh the page to view the results.",
+  "渠道 ${name} 余额更新成功！": "Channel ${name} balance updated successfully!",
+  "已更新完毕所有已启用渠道余额！": "The balance of all enabled channels has been updated!",
   "搜索渠道的 ID，名称和密钥 ...": "Search for channel ID, name and key ...",
   "名称": "Name",
   "分组": "Group",
@@ -141,9 +141,9 @@
   "启用": "Enable",
   "编辑": "Edit",
   "添加新的渠道": "Add a new channel",
-  "测试所有通道": "Test all channels",
-  "测试所有已启用通道": "Test all enabled channels",
-  "更新所有已启用通道余额": "Update the balance of all enabled channels",
+  "测试所有渠道": "Test all channels",
+  "测试所有已启用渠道": "Test all enabled channels",
+  "更新所有已启用渠道余额": "Update the balance of all enabled channels",
   "刷新": "Refresh",
   "处理中...": "Processing...",
   "绑定成功！": "Binding succeeded!",
@@ -207,11 +207,11 @@
   "监控设置": "Monitoring Settings",
   "最长响应时间": "Longest Response Time",
   "单位秒": "Unit in seconds",
-  "当运行通道全部测试时": "When all operating channels are tested",
-  "超过此时间将自动禁用通道": "Channels will be automatically disabled if this time is exceeded",
+  "当运行渠道全部测试时": "When all operating channels are tested",
+  "超过此时间将自动禁用渠道": "Channels will be automatically disabled if this time is exceeded",
   "额度提醒阈值": "Quota reminder threshold",
   "低于此额度时将发送邮件提醒用户": "Email will be sent to remind users when the quota is below this",
-  "失败时自动禁用通道": "Automatically disable the channel when it fails",
+  "失败时自动禁用渠道": "Automatically disable the channel when it fails",
   "保存监控设置": "Save Monitoring Settings",
   "额度设置": "Quota Settings",
   "新用户初始额度": "Initial quota for new users",
@@ -405,7 +405,7 @@
   "镜像": "Mirror",
   "请输入镜像站地址，格式为：https://domain.com，可不填，不填则使用渠道默认值": "Please enter the mirror site address, the format is: https://domain.com, it can be left blank, if left blank, the default value of the channel will be used",
   "模型": "Model",
-  "请选择该通道所支持的模型": "Please select the model supported by the channel",
+  "请选择该渠道所支持的模型": "Please select the model supported by the channel",
   "填入基础模型": "Fill in the basic model",
   "填入所有模型": "Fill in all models",
   "清除所有模型": "Clear all models",
@@ -515,7 +515,7 @@
   "请输入自定义渠道的 Base URL": "Please enter the Base URL of the custom channel",
   "Homepage URL 填": "Fill in the Homepage URL",
   "Authorization callback URL 填": "Fill in the Authorization callback URL",
-  "请为通道命名": "Please name the channel",
+  "请为渠道命名": "Please name the channel",
   "此项可选，用于修改请求体中的模型名称，为一个 JSON 字符串，键为请求中模型名称，值为要替换的模型名称，例如：": "This is optional, used to modify the model name in the request body, it's a JSON string, the key is the model name in the request, and the value is the model name to be replaced, for example:",
   "模型重定向": "Model redirection",
   "请输入渠道对应的鉴权密钥": "Please enter the authentication key corresponding to the channel",
diff --git a/main.go b/main.go
index b20c6daf..67a3cd95 100644
--- a/main.go
+++ b/main.go
@@ -6,13 +6,15 @@ import (
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-contrib/sessions/cookie"
 	"github.com/gin-gonic/gin"
+	_ "github.com/joho/godotenv/autoload"
 	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/client"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/controller"
 	"github.com/songquanpeng/one-api/middleware"
 	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/router"
 	"os"
 	"strconv"
@@ -22,29 +24,22 @@ import (
 var buildFS embed.FS
 
 func main() {
+	common.Init()
 	logger.SetupLogger()
-	logger.SysLog(fmt.Sprintf("One API %s started", common.Version))
-	if os.Getenv("GIN_MODE") != "debug" {
+	logger.SysLogf("One API %s started", common.Version)
+
+	if os.Getenv("GIN_MODE") != gin.DebugMode {
 		gin.SetMode(gin.ReleaseMode)
 	}
 	if config.DebugEnabled {
 		logger.SysLog("running in debug mode")
 	}
-	var err error
+
 	// Initialize SQL Database
-	model.DB, err = model.InitDB("SQL_DSN")
-	if err != nil {
-		logger.FatalLog("failed to initialize database: " + err.Error())
-	}
-	if os.Getenv("LOG_SQL_DSN") != "" {
-		logger.SysLog("using secondary database for table logs")
-		model.LOG_DB, err = model.InitDB("LOG_SQL_DSN")
-		if err != nil {
-			logger.FatalLog("failed to initialize secondary database: " + err.Error())
-		}
-	} else {
-		model.LOG_DB = model.DB
-	}
+	model.InitDB()
+	model.InitLogDB()
+
+	var err error
 	err = model.CreateRootAccountIfNeed()
 	if err != nil {
 		logger.FatalLog("database init error: " + err.Error())
@@ -71,7 +66,7 @@ func main() {
 	}
 	if config.MemoryCacheEnabled {
 		logger.SysLog("memory cache enabled")
-		logger.SysError(fmt.Sprintf("sync frequency: %d seconds", config.SyncFrequency))
+		logger.SysLog(fmt.Sprintf("sync frequency: %d seconds", config.SyncFrequency))
 		model.InitChannelCache()
 	}
 	if config.MemoryCacheEnabled {
@@ -94,6 +89,7 @@ func main() {
 		logger.SysLog("metric enabled, will disable channel if too much request failed")
 	}
 	openai.InitTokenEncoders()
+	client.Init()
 
 	// Initialize HTTP server
 	server := gin.New()
@@ -111,6 +107,7 @@ func main() {
 	if port == "" {
 		port = strconv.Itoa(*common.Port)
 	}
+	logger.SysLogf("server started on http://localhost:%s", port)
 	err = server.Run(":" + port)
 	if err != nil {
 		logger.FatalLog("failed to start HTTP server: " + err.Error())
diff --git a/middleware/auth.go b/middleware/auth.go
index 30997efd..e0019838 100644
--- a/middleware/auth.go
+++ b/middleware/auth.go
@@ -1,10 +1,12 @@
 package middleware
 
 import (
+	"fmt"
 	"github.com/gin-contrib/sessions"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/blacklist"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/network"
 	"github.com/songquanpeng/one-api/model"
 	"net/http"
 	"strings"
@@ -43,7 +45,7 @@ func authHelper(c *gin.Context, minRole int) {
 			return
 		}
 	}
-	if status.(int) == common.UserStatusDisabled || blacklist.IsUserBanned(id.(int)) {
+	if status.(int) == model.UserStatusDisabled || blacklist.IsUserBanned(id.(int)) {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
 			"message": "用户已被封禁",
@@ -70,24 +72,25 @@ func authHelper(c *gin.Context, minRole int) {
 
 func UserAuth() func(c *gin.Context) {
 	return func(c *gin.Context) {
-		authHelper(c, common.RoleCommonUser)
+		authHelper(c, model.RoleCommonUser)
 	}
 }
 
 func AdminAuth() func(c *gin.Context) {
 	return func(c *gin.Context) {
-		authHelper(c, common.RoleAdminUser)
+		authHelper(c, model.RoleAdminUser)
 	}
 }
 
 func RootAuth() func(c *gin.Context) {
 	return func(c *gin.Context) {
-		authHelper(c, common.RoleRootUser)
+		authHelper(c, model.RoleRootUser)
 	}
 }
 
 func TokenAuth() func(c *gin.Context) {
 	return func(c *gin.Context) {
+		ctx := c.Request.Context()
 		key := c.Request.Header.Get("Authorization")
 		key = strings.TrimPrefix(key, "Bearer ")
 		key = strings.TrimPrefix(key, "sk-")
@@ -98,6 +101,12 @@ func TokenAuth() func(c *gin.Context) {
 			abortWithMessage(c, http.StatusUnauthorized, err.Error())
 			return
 		}
+		if token.Subnet != nil && *token.Subnet != "" {
+			if !network.IsIpInSubnets(ctx, c.ClientIP(), *token.Subnet) {
+				abortWithMessage(c, http.StatusForbidden, fmt.Sprintf("该令牌只能在指定网段使用：%s，当前 ip：%s", *token.Subnet, c.ClientIP()))
+				return
+			}
+		}
 		userEnabled, err := model.CacheIsUserEnabled(token.UserId)
 		if err != nil {
 			abortWithMessage(c, http.StatusInternalServerError, err.Error())
@@ -107,17 +116,52 @@ func TokenAuth() func(c *gin.Context) {
 			abortWithMessage(c, http.StatusForbidden, "用户已被封禁")
 			return
 		}
-		c.Set("id", token.UserId)
-		c.Set("token_id", token.Id)
-		c.Set("token_name", token.Name)
+		requestModel, err := getRequestModel(c)
+		if err != nil && shouldCheckModel(c) {
+			abortWithMessage(c, http.StatusBadRequest, err.Error())
+			return
+		}
+		c.Set(ctxkey.RequestModel, requestModel)
+		if token.Models != nil && *token.Models != "" {
+			c.Set(ctxkey.AvailableModels, *token.Models)
+			if requestModel != "" && !isModelInList(requestModel, *token.Models) {
+				abortWithMessage(c, http.StatusForbidden, fmt.Sprintf("该令牌无权使用模型：%s", requestModel))
+				return
+			}
+		}
+		c.Set(ctxkey.Id, token.UserId)
+		c.Set(ctxkey.TokenId, token.Id)
+		c.Set(ctxkey.TokenName, token.Name)
 		if len(parts) > 1 {
 			if model.IsAdmin(token.UserId) {
-				c.Set("specific_channel_id", parts[1])
+				c.Set(ctxkey.SpecificChannelId, parts[1])
 			} else {
 				abortWithMessage(c, http.StatusForbidden, "普通用户不支持指定渠道")
 				return
 			}
 		}
+
+		// set channel id for proxy relay
+		if channelId := c.Param("channelid"); channelId != "" {
+			c.Set(ctxkey.SpecificChannelId, channelId)
+		}
+
 		c.Next()
 	}
 }
+
+func shouldCheckModel(c *gin.Context) bool {
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/completions") {
+		return true
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/chat/completions") {
+		return true
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/images") {
+		return true
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/audio") {
+		return true
+	}
+	return false
+}
diff --git a/middleware/distributor.go b/middleware/distributor.go
index e845c2f8..0aceb29d 100644
--- a/middleware/distributor.go
+++ b/middleware/distributor.go
@@ -2,28 +2,27 @@ package middleware
 
 import (
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/relay/channeltype"
 	"net/http"
 	"strconv"
-	"strings"
-
-	"github.com/gin-gonic/gin"
 )
 
 type ModelRequest struct {
-	Model string `json:"model"`
+	Model string `json:"model" form:"model"`
 }
 
 func Distribute() func(c *gin.Context) {
 	return func(c *gin.Context) {
-		userId := c.GetInt("id")
+		userId := c.GetInt(ctxkey.Id)
 		userGroup, _ := model.CacheGetUserGroup(userId)
-		c.Set("group", userGroup)
+		c.Set(ctxkey.Group, userGroup)
 		var requestModel string
 		var channel *model.Channel
-		channelId, ok := c.Get("specific_channel_id")
+		channelId, ok := c.Get(ctxkey.SpecificChannelId)
 		if ok {
 			id, err := strconv.Atoi(channelId.(string))
 			if err != nil {
@@ -35,42 +34,16 @@ func Distribute() func(c *gin.Context) {
 				abortWithMessage(c, http.StatusBadRequest, "无效的渠道 Id")
 				return
 			}
-			if channel.Status != common.ChannelStatusEnabled {
+			if channel.Status != model.ChannelStatusEnabled {
 				abortWithMessage(c, http.StatusForbidden, "该渠道已被禁用")
 				return
 			}
 		} else {
-			// Select a channel for the user
-			var modelRequest ModelRequest
-			err := common.UnmarshalBodyReusable(c, &modelRequest)
+			requestModel = c.GetString(ctxkey.RequestModel)
+			var err error
+			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, requestModel, false)
 			if err != nil {
-				abortWithMessage(c, http.StatusBadRequest, "无效的请求")
-				return
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/moderations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "text-moderation-stable"
-				}
-			}
-			if strings.HasSuffix(c.Request.URL.Path, "embeddings") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = c.Param("model")
-				}
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "dall-e-2"
-				}
-			}
-			if strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") || strings.HasPrefix(c.Request.URL.Path, "/v1/audio/translations") {
-				if modelRequest.Model == "" {
-					modelRequest.Model = "whisper-1"
-				}
-			}
-			requestModel = modelRequest.Model
-			channel, err = model.CacheGetRandomSatisfiedChannel(userGroup, modelRequest.Model, false)
-			if err != nil {
-				message := fmt.Sprintf("当前分组 %s 下对于模型 %s 无可用渠道", userGroup, modelRequest.Model)
+				message := fmt.Sprintf("当前分组 %s 下对于模型 %s 无可用渠道", userGroup, requestModel)
 				if channel != nil {
 					logger.SysError(fmt.Sprintf("渠道不存在：%d", channel.Id))
 					message = "数据库一致性已被破坏，请联系管理员"
@@ -85,28 +58,41 @@ func Distribute() func(c *gin.Context) {
 }
 
 func SetupContextForSelectedChannel(c *gin.Context, channel *model.Channel, modelName string) {
-	c.Set("channel", channel.Type)
-	c.Set("channel_id", channel.Id)
-	c.Set("channel_name", channel.Name)
-	c.Set("model_mapping", channel.GetModelMapping())
-	c.Set("original_model", modelName) // for retry
+	c.Set(ctxkey.Channel, channel.Type)
+	c.Set(ctxkey.ChannelId, channel.Id)
+	c.Set(ctxkey.ChannelName, channel.Name)
+	if channel.SystemPrompt != nil && *channel.SystemPrompt != "" {
+		c.Set(ctxkey.SystemPrompt, *channel.SystemPrompt)
+	}
+	c.Set(ctxkey.ModelMapping, channel.GetModelMapping())
+	c.Set(ctxkey.OriginalModel, modelName) // for retry
 	c.Request.Header.Set("Authorization", fmt.Sprintf("Bearer %s", channel.Key))
-	c.Set("base_url", channel.GetBaseURL())
-	// this is for backward compatibility
-	switch channel.Type {
-	case common.ChannelTypeAzure:
-		c.Set(common.ConfigKeyAPIVersion, channel.Other)
-	case common.ChannelTypeXunfei:
-		c.Set(common.ConfigKeyAPIVersion, channel.Other)
-	case common.ChannelTypeGemini:
-		c.Set(common.ConfigKeyAPIVersion, channel.Other)
-	case common.ChannelTypeAIProxyLibrary:
-		c.Set(common.ConfigKeyLibraryID, channel.Other)
-	case common.ChannelTypeAli:
-		c.Set(common.ConfigKeyPlugin, channel.Other)
-	}
+	c.Set(ctxkey.BaseURL, channel.GetBaseURL())
 	cfg, _ := channel.LoadConfig()
-	for k, v := range cfg {
-		c.Set(common.ConfigKeyPrefix+k, v)
+	// this is for backward compatibility
+	if channel.Other != nil {
+		switch channel.Type {
+		case channeltype.Azure:
+			if cfg.APIVersion == "" {
+				cfg.APIVersion = *channel.Other
+			}
+		case channeltype.Xunfei:
+			if cfg.APIVersion == "" {
+				cfg.APIVersion = *channel.Other
+			}
+		case channeltype.Gemini:
+			if cfg.APIVersion == "" {
+				cfg.APIVersion = *channel.Other
+			}
+		case channeltype.AIProxyLibrary:
+			if cfg.LibraryID == "" {
+				cfg.LibraryID = *channel.Other
+			}
+		case channeltype.Ali:
+			if cfg.Plugin == "" {
+				cfg.Plugin = *channel.Other
+			}
+		}
 	}
+	c.Set(ctxkey.Config, cfg)
 }
diff --git a/middleware/gzip.go b/middleware/gzip.go
new file mode 100644
index 00000000..4d4ce0c2
--- /dev/null
+++ b/middleware/gzip.go
@@ -0,0 +1,27 @@
+package middleware
+
+import (
+	"compress/gzip"
+	"github.com/gin-gonic/gin"
+	"io"
+	"net/http"
+)
+
+func GzipDecodeMiddleware() gin.HandlerFunc {
+	return func(c *gin.Context) {
+		if c.GetHeader("Content-Encoding") == "gzip" {
+			gzipReader, err := gzip.NewReader(c.Request.Body)
+			if err != nil {
+				c.AbortWithStatus(http.StatusBadRequest)
+				return
+			}
+			defer gzipReader.Close()
+
+			// Replace the request body with the decompressed data
+			c.Request.Body = io.NopCloser(gzipReader)
+		}
+
+		// Continue processing the request
+		c.Next()
+	}
+}
diff --git a/middleware/logger.go b/middleware/logger.go
index 6aae4f23..191364f8 100644
--- a/middleware/logger.go
+++ b/middleware/logger.go
@@ -3,14 +3,14 @@ package middleware
 import (
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/helper"
 )
 
 func SetUpLogger(server *gin.Engine) {
 	server.Use(gin.LoggerWithFormatter(func(param gin.LogFormatterParams) string {
 		var requestID string
 		if param.Keys != nil {
-			requestID = param.Keys[logger.RequestIdKey].(string)
+			requestID = param.Keys[helper.RequestIdKey].(string)
 		}
 		return fmt.Sprintf("[GIN] %s | %s | %3d | %13v | %15s | %7s %s\n",
 			param.TimeStamp.Format("2006/01/02 - 15:04:05"),
diff --git a/middleware/rate-limit.go b/middleware/rate-limit.go
index 0f300f2b..c1be92f3 100644
--- a/middleware/rate-limit.go
+++ b/middleware/rate-limit.go
@@ -3,11 +3,12 @@ package middleware
 import (
 	"context"
 	"fmt"
+	"net/http"
+	"time"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
-	"net/http"
-	"time"
 )
 
 var timeFormat = "2006-01-02T15:04:05.000Z"
@@ -70,6 +71,11 @@ func memoryRateLimiter(c *gin.Context, maxRequestNum int, duration int64, mark s
 }
 
 func rateLimitFactory(maxRequestNum int, duration int64, mark string) func(c *gin.Context) {
+	if maxRequestNum == 0 {
+		return func(c *gin.Context) {
+			c.Next()
+		}
+	}
 	if common.RedisEnabled {
 		return func(c *gin.Context) {
 			redisRateLimiter(c, maxRequestNum, duration, mark)
diff --git a/middleware/request-id.go b/middleware/request-id.go
index a4c49ddb..bef09e32 100644
--- a/middleware/request-id.go
+++ b/middleware/request-id.go
@@ -4,16 +4,15 @@ import (
 	"context"
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
 )
 
 func RequestId() func(c *gin.Context) {
 	return func(c *gin.Context) {
 		id := helper.GenRequestID()
-		c.Set(logger.RequestIdKey, id)
-		ctx := context.WithValue(c.Request.Context(), logger.RequestIdKey, id)
+		c.Set(helper.RequestIdKey, id)
+		ctx := context.WithValue(c.Request.Context(), helper.RequestIdKey, id)
 		c.Request = c.Request.WithContext(ctx)
-		c.Header(logger.RequestIdKey, id)
+		c.Header(helper.RequestIdKey, id)
 		c.Next()
 	}
 }
diff --git a/middleware/utils.go b/middleware/utils.go
index bc14c367..4d2f8092 100644
--- a/middleware/utils.go
+++ b/middleware/utils.go
@@ -1,18 +1,60 @@
 package middleware
 
 import (
+	"fmt"
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
+	"strings"
 )
 
 func abortWithMessage(c *gin.Context, statusCode int, message string) {
 	c.JSON(statusCode, gin.H{
 		"error": gin.H{
-			"message": helper.MessageWithRequestId(message, c.GetString(logger.RequestIdKey)),
+			"message": helper.MessageWithRequestId(message, c.GetString(helper.RequestIdKey)),
 			"type":    "one_api_error",
 		},
 	})
 	c.Abort()
 	logger.Error(c.Request.Context(), message)
 }
+
+func getRequestModel(c *gin.Context) (string, error) {
+	var modelRequest ModelRequest
+	err := common.UnmarshalBodyReusable(c, &modelRequest)
+	if err != nil {
+		return "", fmt.Errorf("common.UnmarshalBodyReusable failed: %w", err)
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/moderations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "text-moderation-stable"
+		}
+	}
+	if strings.HasSuffix(c.Request.URL.Path, "embeddings") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = c.Param("model")
+		}
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "dall-e-2"
+		}
+	}
+	if strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") || strings.HasPrefix(c.Request.URL.Path, "/v1/audio/translations") {
+		if modelRequest.Model == "" {
+			modelRequest.Model = "whisper-1"
+		}
+	}
+	return modelRequest.Model, nil
+}
+
+func isModelInList(modelName string, models string) bool {
+	modelList := strings.Split(models, ",")
+	for _, model := range modelList {
+		if modelName == model {
+			return true
+		}
+	}
+	return false
+}
diff --git a/model/ability.go b/model/ability.go
index 7127abc3..2db72518 100644
--- a/model/ability.go
+++ b/model/ability.go
@@ -1,7 +1,10 @@
 package model
 
 import (
+	"context"
 	"github.com/songquanpeng/one-api/common"
+	"gorm.io/gorm"
+	"sort"
 	"strings"
 )
 
@@ -13,7 +16,7 @@ type Ability struct {
 	Priority  *int64 `json:"priority" gorm:"bigint;default:0;index"`
 }
 
-func GetRandomSatisfiedChannel(group string, model string) (*Channel, error) {
+func GetRandomSatisfiedChannel(group string, model string, ignoreFirstPriority bool) (*Channel, error) {
 	ability := Ability{}
 	groupCol := "`group`"
 	trueVal := "1"
@@ -23,8 +26,13 @@ func GetRandomSatisfiedChannel(group string, model string) (*Channel, error) {
 	}
 
 	var err error = nil
-	maxPrioritySubQuery := DB.Model(&Ability{}).Select("MAX(priority)").Where(groupCol+" = ? and model = ? and enabled = "+trueVal, group, model)
-	channelQuery := DB.Where(groupCol+" = ? and model = ? and enabled = "+trueVal+" and priority = (?)", group, model, maxPrioritySubQuery)
+	var channelQuery *gorm.DB
+	if ignoreFirstPriority {
+		channelQuery = DB.Where(groupCol+" = ? and model = ? and enabled = "+trueVal, group, model)
+	} else {
+		maxPrioritySubQuery := DB.Model(&Ability{}).Select("MAX(priority)").Where(groupCol+" = ? and model = ? and enabled = "+trueVal, group, model)
+		channelQuery = DB.Where(groupCol+" = ? and model = ? and enabled = "+trueVal+" and priority = (?)", group, model, maxPrioritySubQuery)
+	}
 	if common.UsingSQLite || common.UsingPostgreSQL {
 		err = channelQuery.Order("RANDOM()").First(&ability).Error
 	} else {
@@ -49,7 +57,7 @@ func (channel *Channel) AddAbilities() error {
 				Group:     group,
 				Model:     model,
 				ChannelId: channel.Id,
-				Enabled:   channel.Status == common.ChannelStatusEnabled,
+				Enabled:   channel.Status == ChannelStatusEnabled,
 				Priority:  channel.Priority,
 			}
 			abilities = append(abilities, ability)
@@ -82,3 +90,19 @@ func (channel *Channel) UpdateAbilities() error {
 func UpdateAbilityStatus(channelId int, status bool) error {
 	return DB.Model(&Ability{}).Where("channel_id = ?", channelId).Select("enabled").Update("enabled", status).Error
 }
+
+func GetGroupModels(ctx context.Context, group string) ([]string, error) {
+	groupCol := "`group`"
+	trueVal := "1"
+	if common.UsingPostgreSQL {
+		groupCol = `"group"`
+		trueVal = "true"
+	}
+	var models []string
+	err := DB.Model(&Ability{}).Distinct("model").Where(groupCol+" = ? and enabled = "+trueVal, group).Pluck("model", &models).Error
+	if err != nil {
+		return nil, err
+	}
+	sort.Strings(models)
+	return models, err
+}
diff --git a/model/cache.go b/model/cache.go
index dd20d857..cfb0f8a4 100644
--- a/model/cache.go
+++ b/model/cache.go
@@ -8,6 +8,7 @@ import (
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
 	"math/rand"
 	"sort"
 	"strconv"
@@ -21,6 +22,7 @@ var (
 	UserId2GroupCacheSeconds  = config.SyncFrequency
 	UserId2QuotaCacheSeconds  = config.SyncFrequency
 	UserId2StatusCacheSeconds = config.SyncFrequency
+	GroupModelsCacheSeconds   = config.SyncFrequency
 )
 
 func CacheGetTokenByKey(key string) (*Token, error) {
@@ -146,13 +148,32 @@ func CacheIsUserEnabled(userId int) (bool, error) {
 	return userEnabled, err
 }
 
+func CacheGetGroupModels(ctx context.Context, group string) ([]string, error) {
+	if !common.RedisEnabled {
+		return GetGroupModels(ctx, group)
+	}
+	modelsStr, err := common.RedisGet(fmt.Sprintf("group_models:%s", group))
+	if err == nil {
+		return strings.Split(modelsStr, ","), nil
+	}
+	models, err := GetGroupModels(ctx, group)
+	if err != nil {
+		return nil, err
+	}
+	err = common.RedisSet(fmt.Sprintf("group_models:%s", group), strings.Join(models, ","), time.Duration(GroupModelsCacheSeconds)*time.Second)
+	if err != nil {
+		logger.SysError("Redis set group models error: " + err.Error())
+	}
+	return models, nil
+}
+
 var group2model2channels map[string]map[string][]*Channel
 var channelSyncLock sync.RWMutex
 
 func InitChannelCache() {
 	newChannelId2channel := make(map[int]*Channel)
 	var channels []*Channel
-	DB.Where("status = ?", common.ChannelStatusEnabled).Find(&channels)
+	DB.Where("status = ?", ChannelStatusEnabled).Find(&channels)
 	for _, channel := range channels {
 		newChannelId2channel[channel.Id] = channel
 	}
@@ -205,7 +226,7 @@ func SyncChannelCache(frequency int) {
 
 func CacheGetRandomSatisfiedChannel(group string, model string, ignoreFirstPriority bool) (*Channel, error) {
 	if !config.MemoryCacheEnabled {
-		return GetRandomSatisfiedChannel(group, model)
+		return GetRandomSatisfiedChannel(group, model, ignoreFirstPriority)
 	}
 	channelSyncLock.RLock()
 	defer channelSyncLock.RUnlock()
@@ -227,7 +248,7 @@ func CacheGetRandomSatisfiedChannel(group string, model string, ignoreFirstPrior
 	idx := rand.Intn(endIdx)
 	if ignoreFirstPriority {
 		if endIdx < len(channels) { // which means there are more than one priority
-			idx = common.RandRange(endIdx, len(channels))
+			idx = random.RandRange(endIdx, len(channels))
 		}
 	}
 	return channels[idx], nil
diff --git a/model/channel.go b/model/channel.go
index fc4905b1..4b0f4b01 100644
--- a/model/channel.go
+++ b/model/channel.go
@@ -3,13 +3,20 @@ package model
 import (
 	"encoding/json"
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
+
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
 	"gorm.io/gorm"
 )
 
+const (
+	ChannelStatusUnknown          = 0
+	ChannelStatusEnabled          = 1 // don't use 0, 0 is the default value!
+	ChannelStatusManuallyDisabled = 2 // also don't use 0
+	ChannelStatusAutoDisabled     = 3
+)
+
 type Channel struct {
 	Id                 int     `json:"id"`
 	Type               int     `json:"type" gorm:"default:0"`
@@ -21,7 +28,7 @@ type Channel struct {
 	TestTime           int64   `json:"test_time" gorm:"bigint"`
 	ResponseTime       int     `json:"response_time"` // in milliseconds
 	BaseURL            *string `json:"base_url" gorm:"column:base_url;default:''"`
-	Other              string  `json:"other"`   // DEPRECATED: please save config to field Config
+	Other              *string `json:"other"`   // DEPRECATED: please save config to field Config
 	Balance            float64 `json:"balance"` // in USD
 	BalanceUpdatedTime int64   `json:"balance_updated_time" gorm:"bigint"`
 	Models             string  `json:"models"`
@@ -30,6 +37,19 @@ type Channel struct {
 	ModelMapping       *string `json:"model_mapping" gorm:"type:varchar(1024);default:''"`
 	Priority           *int64  `json:"priority" gorm:"bigint;default:0"`
 	Config             string  `json:"config"`
+	SystemPrompt       *string `json:"system_prompt" gorm:"type:text"`
+}
+
+type ChannelConfig struct {
+	Region            string `json:"region,omitempty"`
+	SK                string `json:"sk,omitempty"`
+	AK                string `json:"ak,omitempty"`
+	UserID            string `json:"user_id,omitempty"`
+	APIVersion        string `json:"api_version,omitempty"`
+	LibraryID         string `json:"library_id,omitempty"`
+	Plugin            string `json:"plugin,omitempty"`
+	VertexAIProjectID string `json:"vertex_ai_project_id,omitempty"`
+	VertexAIADC       string `json:"vertex_ai_adc,omitempty"`
 }
 
 func GetAllChannels(startIdx int, num int, scope string) ([]*Channel, error) {
@@ -39,7 +59,7 @@ func GetAllChannels(startIdx int, num int, scope string) ([]*Channel, error) {
 	case "all":
 		err = DB.Order("id desc").Find(&channels).Error
 	case "disabled":
-		err = DB.Order("id desc").Where("status = ? or status = ?", common.ChannelStatusAutoDisabled, common.ChannelStatusManuallyDisabled).Find(&channels).Error
+		err = DB.Order("id desc").Where("status = ? or status = ?", ChannelStatusAutoDisabled, ChannelStatusManuallyDisabled).Find(&channels).Error
 	default:
 		err = DB.Order("id desc").Limit(num).Offset(startIdx).Omit("key").Find(&channels).Error
 	}
@@ -155,20 +175,20 @@ func (channel *Channel) Delete() error {
 	return err
 }
 
-func (channel *Channel) LoadConfig() (map[string]string, error) {
+func (channel *Channel) LoadConfig() (ChannelConfig, error) {
+	var cfg ChannelConfig
 	if channel.Config == "" {
-		return nil, nil
+		return cfg, nil
 	}
-	cfg := make(map[string]string)
 	err := json.Unmarshal([]byte(channel.Config), &cfg)
 	if err != nil {
-		return nil, err
+		return cfg, err
 	}
 	return cfg, nil
 }
 
 func UpdateChannelStatusById(id int, status int) {
-	err := UpdateAbilityStatus(id, status == common.ChannelStatusEnabled)
+	err := UpdateAbilityStatus(id, status == ChannelStatusEnabled)
 	if err != nil {
 		logger.SysError("failed to update ability status: " + err.Error())
 	}
@@ -199,6 +219,6 @@ func DeleteChannelByStatus(status int64) (int64, error) {
 }
 
 func DeleteDisabledChannel() (int64, error) {
-	result := DB.Where("status = ? or status = ?", common.ChannelStatusAutoDisabled, common.ChannelStatusManuallyDisabled).Delete(&Channel{})
+	result := DB.Where("status = ? or status = ?", ChannelStatusAutoDisabled, ChannelStatusManuallyDisabled).Delete(&Channel{})
 	return result.RowsAffected, result.Error
 }
diff --git a/model/log.go b/model/log.go
index 4409f73e..58fdd513 100644
--- a/model/log.go
+++ b/model/log.go
@@ -3,11 +3,11 @@ package model
 import (
 	"context"
 	"fmt"
+
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
-
 	"gorm.io/gorm"
 )
 
@@ -51,6 +51,21 @@ func RecordLog(userId int, logType int, content string) {
 	}
 }
 
+func RecordTopupLog(userId int, content string, quota int) {
+	log := &Log{
+		UserId:    userId,
+		Username:  GetUsernameById(userId),
+		CreatedAt: helper.GetTimestamp(),
+		Type:      LogTypeTopup,
+		Content:   content,
+		Quota:     quota,
+	}
+	err := LOG_DB.Create(log).Error
+	if err != nil {
+		logger.SysError("failed to record log: " + err.Error())
+	}
+}
+
 func RecordConsumeLog(ctx context.Context, userId int, channelId int, promptTokens int, completionTokens int, modelName string, tokenName string, quota int64, content string) {
 	logger.Info(ctx, fmt.Sprintf("record consume log: userId=%d, channelId=%d, promptTokens=%d, completionTokens=%d, modelName=%s, tokenName=%s, quota=%d, content=%s", userId, channelId, promptTokens, completionTokens, modelName, tokenName, quota, content))
 	if !config.LogConsumeEnabled {
@@ -138,7 +153,11 @@ func SearchUserLogs(userId int, keyword string) (logs []*Log, err error) {
 }
 
 func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string, channel int) (quota int64) {
-	tx := LOG_DB.Table("logs").Select("ifnull(sum(quota),0)")
+	ifnull := "ifnull"
+	if common.UsingPostgreSQL {
+		ifnull = "COALESCE"
+	}
+	tx := LOG_DB.Table("logs").Select(fmt.Sprintf("%s(sum(quota),0)", ifnull))
 	if username != "" {
 		tx = tx.Where("username = ?", username)
 	}
@@ -162,7 +181,11 @@ func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelNa
 }
 
 func SumUsedToken(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string) (token int) {
-	tx := LOG_DB.Table("logs").Select("ifnull(sum(prompt_tokens),0) + ifnull(sum(completion_tokens),0)")
+	ifnull := "ifnull"
+	if common.UsingPostgreSQL {
+		ifnull = "COALESCE"
+	}
+	tx := LOG_DB.Table("logs").Select(fmt.Sprintf("%s(sum(prompt_tokens),0) + %s(sum(completion_tokens),0)", ifnull, ifnull))
 	if username != "" {
 		tx = tx.Where("username = ?", username)
 	}
diff --git a/model/main.go b/model/main.go
index 60b9d849..6d567e17 100644
--- a/model/main.go
+++ b/model/main.go
@@ -1,12 +1,14 @@
 package model
 
 import (
+	"database/sql"
 	"fmt"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/env"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
 	"gorm.io/driver/mysql"
 	"gorm.io/driver/postgres"
 	"gorm.io/driver/sqlite"
@@ -28,13 +30,17 @@ func CreateRootAccountIfNeed() error {
 		if err != nil {
 			return err
 		}
+		accessToken := random.GetUUID()
+		if config.InitialRootAccessToken != "" {
+			accessToken = config.InitialRootAccessToken
+		}
 		rootUser := User{
 			Username:    "root",
 			Password:    hashedPassword,
-			Role:        common.RoleRootUser,
-			Status:      common.UserStatusEnabled,
+			Role:        RoleRootUser,
+			Status:      UserStatusEnabled,
 			DisplayName: "Root User",
-			AccessToken: helper.GetUUID(),
+			AccessToken: accessToken,
 			Quota:       500000000000000,
 		}
 		DB.Create(&rootUser)
@@ -44,7 +50,7 @@ func CreateRootAccountIfNeed() error {
 				Id:             1,
 				UserId:         rootUser.Id,
 				Key:            config.InitialRootToken,
-				Status:         common.TokenStatusEnabled,
+				Status:         TokenStatusEnabled,
 				Name:           "Initial Root Token",
 				CreatedTime:    helper.GetTimestamp(),
 				AccessedTime:   helper.GetTimestamp(),
@@ -59,92 +65,153 @@ func CreateRootAccountIfNeed() error {
 }
 
 func chooseDB(envName string) (*gorm.DB, error) {
-	if os.Getenv(envName) != "" {
-		dsn := os.Getenv(envName)
-		if strings.HasPrefix(dsn, "postgres://") {
-			// Use PostgreSQL
-			logger.SysLog("using PostgreSQL as database")
-			common.UsingPostgreSQL = true
-			return gorm.Open(postgres.New(postgres.Config{
-				DSN:                  dsn,
-				PreferSimpleProtocol: true, // disables implicit prepared statement usage
-			}), &gorm.Config{
-				PrepareStmt: true, // precompile SQL
-			})
-		}
+	dsn := os.Getenv(envName)
+
+	switch {
+	case strings.HasPrefix(dsn, "postgres://"):
+		// Use PostgreSQL
+		return openPostgreSQL(dsn)
+	case dsn != "":
 		// Use MySQL
-		logger.SysLog("using MySQL as database")
-		common.UsingMySQL = true
-		return gorm.Open(mysql.Open(dsn), &gorm.Config{
-			PrepareStmt: true, // precompile SQL
-		})
+		return openMySQL(dsn)
+	default:
+		// Use SQLite
+		return openSQLite()
 	}
-	// Use SQLite
-	logger.SysLog("SQL_DSN not set, using SQLite as database")
-	common.UsingSQLite = true
-	config := fmt.Sprintf("?_busy_timeout=%d", common.SQLiteBusyTimeout)
-	return gorm.Open(sqlite.Open(common.SQLitePath+config), &gorm.Config{
+}
+
+func openPostgreSQL(dsn string) (*gorm.DB, error) {
+	logger.SysLog("using PostgreSQL as database")
+	common.UsingPostgreSQL = true
+	return gorm.Open(postgres.New(postgres.Config{
+		DSN:                  dsn,
+		PreferSimpleProtocol: true, // disables implicit prepared statement usage
+	}), &gorm.Config{
 		PrepareStmt: true, // precompile SQL
 	})
 }
 
-func InitDB(envName string) (db *gorm.DB, err error) {
-	db, err = chooseDB(envName)
-	if err == nil {
-		if config.DebugSQLEnabled {
-			db = db.Debug()
-		}
-		sqlDB, err := db.DB()
-		if err != nil {
-			return nil, err
-		}
-		sqlDB.SetMaxIdleConns(env.Int("SQL_MAX_IDLE_CONNS", 100))
-		sqlDB.SetMaxOpenConns(env.Int("SQL_MAX_OPEN_CONNS", 1000))
-		sqlDB.SetConnMaxLifetime(time.Second * time.Duration(env.Int("SQL_MAX_LIFETIME", 60)))
+func openMySQL(dsn string) (*gorm.DB, error) {
+	logger.SysLog("using MySQL as database")
+	common.UsingMySQL = true
+	return gorm.Open(mysql.Open(dsn), &gorm.Config{
+		PrepareStmt: true, // precompile SQL
+	})
+}
 
-		if !config.IsMasterNode {
-			return db, err
-		}
-		if common.UsingMySQL {
-			_, _ = sqlDB.Exec("DROP INDEX idx_channels_key ON channels;") // TODO: delete this line when most users have upgraded
-		}
-		if env.Bool("StartSqlMigration", false) {
-			logger.SysLog("database migration started")
-			err = db.AutoMigrate(&Channel{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&Token{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&User{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&Option{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&Redemption{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&Ability{})
-			if err != nil {
-				return nil, err
-			}
-			err = db.AutoMigrate(&Log{})
-			if err != nil {
-				return nil, err
-			}
-			logger.SysLog("database migrated")
-		}
-		return db, err
-	} else {
-		logger.FatalLog(err)
+func openSQLite() (*gorm.DB, error) {
+	logger.SysLog("SQL_DSN not set, using SQLite as database")
+	common.UsingSQLite = true
+	dsn := fmt.Sprintf("%s?_busy_timeout=%d", common.SQLitePath, common.SQLiteBusyTimeout)
+	return gorm.Open(sqlite.Open(dsn), &gorm.Config{
+		PrepareStmt: true, // precompile SQL
+	})
+}
+
+func InitDB() {
+	var err error
+	DB, err = chooseDB("SQL_DSN")
+	if err != nil {
+		logger.FatalLog("failed to initialize database: " + err.Error())
+		return
 	}
-	return db, err
+
+
+	}
+
+	if common.UsingMySQL {
+		_, _ = sqlDB.Exec("DROP INDEX idx_channels_key ON channels;") // TODO: delete this line when most users have upgraded
+	}
+
+	logger.SysLog("database migration started")
+	if err = migrateDB(); err != nil {
+		logger.FatalLog("failed to migrate database: " + err.Error())
+		return
+	}
+	logger.SysLog("database migrated")
+}
+
+func migrateDB() error {
+	var err error
+	if err = DB.AutoMigrate(&Channel{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Token{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&User{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Option{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Redemption{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Ability{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Log{}); err != nil {
+		return err
+	}
+	if err = DB.AutoMigrate(&Channel{}); err != nil {
+		return err
+	}
+	return nil
+}
+
+func InitLogDB() {
+	if os.Getenv("LOG_SQL_DSN") == "" {
+		LOG_DB = DB
+		return
+	}
+
+	logger.SysLog("using secondary database for table logs")
+	var err error
+	LOG_DB, err = chooseDB("LOG_SQL_DSN")
+	if err != nil {
+		logger.FatalLog("failed to initialize secondary database: " + err.Error())
+		return
+	}
+
+	setDBConns(LOG_DB)
+
+	if !config.IsMasterNode {
+		return
+	}
+
+	logger.SysLog("secondary database migration started")
+	err = migrateLOGDB()
+	if err != nil {
+		logger.FatalLog("failed to migrate secondary database: " + err.Error())
+		return
+	}
+	logger.SysLog("secondary database migrated")
+}
+
+func migrateLOGDB() error {
+	var err error
+	if err = LOG_DB.AutoMigrate(&Log{}); err != nil {
+		return err
+	}
+	return nil
+}
+
+func setDBConns(db *gorm.DB) *sql.DB {
+	if config.DebugSQLEnabled {
+		db = db.Debug()
+	}
+
+	sqlDB, err := db.DB()
+	if err != nil {
+		logger.FatalLog("failed to connect database: " + err.Error())
+		return nil
+	}
+
+	sqlDB.SetMaxIdleConns(env.Int("SQL_MAX_IDLE_CONNS", 100))
+	sqlDB.SetMaxOpenConns(env.Int("SQL_MAX_OPEN_CONNS", 1000))
+	sqlDB.SetConnMaxLifetime(time.Second * time.Duration(env.Int("SQL_MAX_LIFETIME", 60)))
+	return sqlDB
 }
 
 func closeDB(db *gorm.DB) error {
diff --git a/model/option.go b/model/option.go
index 1d1c28b4..8fd30aee 100644
--- a/model/option.go
+++ b/model/option.go
@@ -1,9 +1,9 @@
 package model
 
 import (
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/logger"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
 	"strconv"
 	"strings"
 	"time"
@@ -28,6 +28,7 @@ func InitOptionMap() {
 	config.OptionMap["PasswordRegisterEnabled"] = strconv.FormatBool(config.PasswordRegisterEnabled)
 	config.OptionMap["EmailVerificationEnabled"] = strconv.FormatBool(config.EmailVerificationEnabled)
 	config.OptionMap["GitHubOAuthEnabled"] = strconv.FormatBool(config.GitHubOAuthEnabled)
+	config.OptionMap["OidcEnabled"] = strconv.FormatBool(config.OidcEnabled)
 	config.OptionMap["WeChatAuthEnabled"] = strconv.FormatBool(config.WeChatAuthEnabled)
 	config.OptionMap["TurnstileCheckEnabled"] = strconv.FormatBool(config.TurnstileCheckEnabled)
 	config.OptionMap["RegisterEnabled"] = strconv.FormatBool(config.RegisterEnabled)
@@ -66,9 +67,9 @@ func InitOptionMap() {
 	config.OptionMap["QuotaForInvitee"] = strconv.FormatInt(config.QuotaForInvitee, 10)
 	config.OptionMap["QuotaRemindThreshold"] = strconv.FormatInt(config.QuotaRemindThreshold, 10)
 	config.OptionMap["PreConsumedQuota"] = strconv.FormatInt(config.PreConsumedQuota, 10)
-	config.OptionMap["ModelRatio"] = common.ModelRatio2JSONString()
-	config.OptionMap["GroupRatio"] = common.GroupRatio2JSONString()
-	config.OptionMap["CompletionRatio"] = common.CompletionRatio2JSONString()
+	config.OptionMap["ModelRatio"] = billingratio.ModelRatio2JSONString()
+	config.OptionMap["GroupRatio"] = billingratio.GroupRatio2JSONString()
+	config.OptionMap["CompletionRatio"] = billingratio.CompletionRatio2JSONString()
 	config.OptionMap["TopUpLink"] = config.TopUpLink
 	config.OptionMap["ChatLink"] = config.ChatLink
 	config.OptionMap["QuotaPerUnit"] = strconv.FormatFloat(config.QuotaPerUnit, 'f', -1, 64)
@@ -82,7 +83,7 @@ func loadOptionsFromDatabase() {
 	options, _ := AllOption()
 	for _, option := range options {
 		if option.Key == "ModelRatio" {
-			option.Value = common.AddNewMissingRatio(option.Value)
+			option.Value = billingratio.AddNewMissingRatio(option.Value)
 		}
 		err := updateOptionMap(option.Key, option.Value)
 		if err != nil {
@@ -130,6 +131,8 @@ func updateOptionMap(key string, value string) (err error) {
 			config.EmailVerificationEnabled = boolValue
 		case "GitHubOAuthEnabled":
 			config.GitHubOAuthEnabled = boolValue
+		case "OidcEnabled":
+			config.OidcEnabled = boolValue
 		case "WeChatAuthEnabled":
 			config.WeChatAuthEnabled = boolValue
 		case "TurnstileCheckEnabled":
@@ -172,6 +175,22 @@ func updateOptionMap(key string, value string) (err error) {
 		config.GitHubClientId = value
 	case "GitHubClientSecret":
 		config.GitHubClientSecret = value
+	case "LarkClientId":
+		config.LarkClientId = value
+	case "LarkClientSecret":
+		config.LarkClientSecret = value
+	case "OidcClientId":
+		config.OidcClientId = value
+	case "OidcClientSecret":
+		config.OidcClientSecret = value
+	case "OidcWellKnown":
+		config.OidcWellKnown = value
+	case "OidcAuthorizationEndpoint":
+		config.OidcAuthorizationEndpoint = value
+	case "OidcTokenEndpoint":
+		config.OidcTokenEndpoint = value
+	case "OidcUserinfoEndpoint":
+		config.OidcUserinfoEndpoint = value
 	case "Footer":
 		config.Footer = value
 	case "SystemName":
@@ -205,11 +224,11 @@ func updateOptionMap(key string, value string) (err error) {
 	case "RetryTimes":
 		config.RetryTimes, _ = strconv.Atoi(value)
 	case "ModelRatio":
-		err = common.UpdateModelRatioByJSONString(value)
+		err = billingratio.UpdateModelRatioByJSONString(value)
 	case "GroupRatio":
-		err = common.UpdateGroupRatioByJSONString(value)
+		err = billingratio.UpdateGroupRatioByJSONString(value)
 	case "CompletionRatio":
-		err = common.UpdateCompletionRatioByJSONString(value)
+		err = billingratio.UpdateCompletionRatioByJSONString(value)
 	case "TopUpLink":
 		config.TopUpLink = value
 	case "ChatLink":
diff --git a/model/redemption.go b/model/redemption.go
index 79a5b8a9..45871a71 100644
--- a/model/redemption.go
+++ b/model/redemption.go
@@ -8,6 +8,12 @@ import (
 	"gorm.io/gorm"
 )
 
+const (
+	RedemptionCodeStatusEnabled  = 1 // don't use 0, 0 is the default value!
+	RedemptionCodeStatusDisabled = 2 // also don't use 0
+	RedemptionCodeStatusUsed     = 3 // also don't use 0
+)
+
 type Redemption struct {
 	Id           int    `json:"id"`
 	UserId       int    `json:"user_id"`
@@ -61,7 +67,7 @@ func Redeem(key string, userId int) (quota int64, err error) {
 		if err != nil {
 			return errors.New("无效的兑换码")
 		}
-		if redemption.Status != common.RedemptionCodeStatusEnabled {
+		if redemption.Status != RedemptionCodeStatusEnabled {
 			return errors.New("该兑换码已被使用")
 		}
 		err = tx.Model(&User{}).Where("id = ?", userId).Update("quota", gorm.Expr("quota + ?", redemption.Quota)).Error
@@ -69,7 +75,7 @@ func Redeem(key string, userId int) (quota int64, err error) {
 			return err
 		}
 		redemption.RedeemedTime = helper.GetTimestamp()
-		redemption.Status = common.RedemptionCodeStatusUsed
+		redemption.Status = RedemptionCodeStatusUsed
 		err = tx.Save(redemption).Error
 		return err
 	})
diff --git a/model/token.go b/model/token.go
index 493e27c9..91e72a82 100644
--- a/model/token.go
+++ b/model/token.go
@@ -11,25 +11,34 @@ import (
 	"gorm.io/gorm"
 )
 
+const (
+	TokenStatusEnabled   = 1 // don't use 0, 0 is the default value!
+	TokenStatusDisabled  = 2 // also don't use 0
+	TokenStatusExpired   = 3
+	TokenStatusExhausted = 4
+)
+
 type Token struct {
-	Id             int    `json:"id"`
-	UserId         int    `json:"user_id"`
-	Key            string `json:"key" gorm:"type:char(48);uniqueIndex"`
-	Status         int    `json:"status" gorm:"default:1"`
-	Name           string `json:"name" gorm:"index" `
-	CreatedTime    int64  `json:"created_time" gorm:"bigint"`
-	AccessedTime   int64  `json:"accessed_time" gorm:"bigint"`
-	ExpiredTime    int64  `json:"expired_time" gorm:"bigint;default:-1"` // -1 means never expired
-	RemainQuota    int64  `json:"remain_quota" gorm:"bigint;default:0"`
-	UnlimitedQuota bool   `json:"unlimited_quota" gorm:"default:false"`
-	UsedQuota      int64  `json:"used_quota" gorm:"bigint;default:0"` // used quota
+	Id             int     `json:"id"`
+	UserId         int     `json:"user_id"`
+	Key            string  `json:"key" gorm:"type:char(48);uniqueIndex"`
+	Status         int     `json:"status" gorm:"default:1"`
+	Name           string  `json:"name" gorm:"index" `
+	CreatedTime    int64   `json:"created_time" gorm:"bigint"`
+	AccessedTime   int64   `json:"accessed_time" gorm:"bigint"`
+	ExpiredTime    int64   `json:"expired_time" gorm:"bigint;default:-1"` // -1 means never expired
+	RemainQuota    int64   `json:"remain_quota" gorm:"bigint;default:0"`
+	UnlimitedQuota bool    `json:"unlimited_quota" gorm:"default:false"`
+	UsedQuota      int64   `json:"used_quota" gorm:"bigint;default:0"` // used quota
+	Models         *string `json:"models" gorm:"type:text"`            // allowed models
+	Subnet         *string `json:"subnet" gorm:"default:''"`           // allowed subnet
 }
 
 func GetAllUserTokens(userId int, startIdx int, num int, order string) ([]*Token, error) {
 	var tokens []*Token
 	var err error
 	query := DB.Where("user_id = ?", userId)
-	
+
 	switch order {
 	case "remain_quota":
 		query = query.Order("unlimited_quota desc, remain_quota desc")
@@ -38,7 +47,7 @@ func GetAllUserTokens(userId int, startIdx int, num int, order string) ([]*Token
 	default:
 		query = query.Order("id desc")
 	}
-	
+
 	err = query.Limit(num).Offset(startIdx).Find(&tokens).Error
 	return tokens, err
 }
@@ -60,17 +69,17 @@ func ValidateUserToken(key string) (token *Token, err error) {
 		}
 		return nil, errors.New("令牌验证失败")
 	}
-	if token.Status == common.TokenStatusExhausted {
-		return nil, errors.New("该令牌额度已用尽")
-	} else if token.Status == common.TokenStatusExpired {
+	if token.Status == TokenStatusExhausted {
+		return nil, fmt.Errorf("令牌 %s（#%d）额度已用尽", token.Name, token.Id)
+	} else if token.Status == TokenStatusExpired {
 		return nil, errors.New("该令牌已过期")
 	}
-	if token.Status != common.TokenStatusEnabled {
+	if token.Status != TokenStatusEnabled {
 		return nil, errors.New("该令牌状态不可用")
 	}
 	if token.ExpiredTime != -1 && token.ExpiredTime < helper.GetTimestamp() {
 		if !common.RedisEnabled {
-			token.Status = common.TokenStatusExpired
+			token.Status = TokenStatusExpired
 			err := token.SelectUpdate()
 			if err != nil {
 				logger.SysError("failed to update token status" + err.Error())
@@ -81,7 +90,7 @@ func ValidateUserToken(key string) (token *Token, err error) {
 	if !token.UnlimitedQuota && token.RemainQuota <= 0 {
 		if !common.RedisEnabled {
 			// in this case, we can make sure the token is exhausted
-			token.Status = common.TokenStatusExhausted
+			token.Status = TokenStatusExhausted
 			err := token.SelectUpdate()
 			if err != nil {
 				logger.SysError("failed to update token status" + err.Error())
@@ -112,30 +121,40 @@ func GetTokenById(id int) (*Token, error) {
 	return &token, err
 }
 
-func (token *Token) Insert() error {
+func (t *Token) Insert() error {
 	var err error
-	err = DB.Create(token).Error
+	err = DB.Create(t).Error
 	return err
 }
 
 // Update Make sure your token's fields is completed, because this will update non-zero values
-func (token *Token) Update() error {
+func (t *Token) Update() error {
 	var err error
-	err = DB.Model(token).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota").Updates(token).Error
+	err = DB.Model(t).Select("name", "status", "expired_time", "remain_quota", "unlimited_quota", "models", "subnet").Updates(t).Error
 	return err
 }
 
-func (token *Token) SelectUpdate() error {
+func (t *Token) SelectUpdate() error {
 	// This can update zero values
-	return DB.Model(token).Select("accessed_time", "status").Updates(token).Error
+	return DB.Model(t).Select("accessed_time", "status").Updates(t).Error
 }
 
-func (token *Token) Delete() error {
+func (t *Token) Delete() error {
 	var err error
-	err = DB.Delete(token).Error
+	err = DB.Delete(t).Error
 	return err
 }
 
+func (t *Token) GetModels() string {
+	if t == nil {
+		return ""
+	}
+	if t.Models == nil {
+		return ""
+	}
+	return *t.Models
+}
+
 func DeleteTokenById(id int, userId int) (err error) {
 	// Why we need userId here? In case user want to delete other's token.
 	if id == 0 || userId == 0 {
@@ -245,14 +264,14 @@ func PreConsumeTokenQuota(tokenId int, quota int64) (err error) {
 
 func PostConsumeTokenQuota(tokenId int, quota int64) (err error) {
 	token, err := GetTokenById(tokenId)
+	if err != nil {
+		return err
+	}
 	if quota > 0 {
 		err = DecreaseUserQuota(token.UserId, quota)
 	} else {
 		err = IncreaseUserQuota(token.UserId, -quota)
 	}
-	if err != nil {
-		return err
-	}
 	if !token.UnlimitedQuota {
 		if quota > 0 {
 			err = DecreaseTokenQuota(tokenId, quota)
diff --git a/model/user.go b/model/user.go
index 5e729b5e..a964a0d7 100644
--- a/model/user.go
+++ b/model/user.go
@@ -8,10 +8,24 @@ import (
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
 	"gorm.io/gorm"
 	"strings"
 )
 
+const (
+	RoleGuestUser  = 0
+	RoleCommonUser = 1
+	RoleAdminUser  = 10
+	RoleRootUser   = 100
+)
+
+const (
+	UserStatusEnabled  = 1 // don't use 0, 0 is the default value!
+	UserStatusDisabled = 2 // also don't use 0
+	UserStatusDeleted  = 3
+)
+
 // User if you add sensitive fields, don't forget to clean them in setupLogin function.
 // Otherwise, the sensitive information will be saved on local storage in plain text!
 type User struct {
@@ -24,6 +38,8 @@ type User struct {
 	Email            string `json:"email" gorm:"index" validate:"max=50"`
 	GitHubId         string `json:"github_id" gorm:"column:github_id;index"`
 	WeChatId         string `json:"wechat_id" gorm:"column:wechat_id;index"`
+	LarkId           string `json:"lark_id" gorm:"column:lark_id;index"`
+	OidcId           string `json:"oidc_id" gorm:"column:oidc_id;index"`
 	VerificationCode string `json:"verification_code" gorm:"-:all"`                                    // this field is only for Email verification, don't save it to database!
 	AccessToken      string `json:"access_token" gorm:"type:char(32);column:access_token;uniqueIndex"` // this token is for system management
 	Quota            int64  `json:"quota" gorm:"bigint;default:0"`
@@ -41,21 +57,21 @@ func GetMaxUserId() int {
 }
 
 func GetAllUsers(startIdx int, num int, order string) (users []*User, err error) {
-    query := DB.Limit(num).Offset(startIdx).Omit("password").Where("status != ?", common.UserStatusDeleted)
-    
-    switch order {
-    case "quota":
-        query = query.Order("quota desc")
-    case "used_quota":
-        query = query.Order("used_quota desc")
-    case "request_count":
-        query = query.Order("request_count desc")
-    default:
-        query = query.Order("id desc")
-    }
-    
-    err = query.Find(&users).Error
-    return users, err
+	query := DB.Limit(num).Offset(startIdx).Omit("password").Where("status != ?", UserStatusDeleted)
+
+	switch order {
+	case "quota":
+		query = query.Order("quota desc")
+	case "used_quota":
+		query = query.Order("used_quota desc")
+	case "request_count":
+		query = query.Order("request_count desc")
+	default:
+		query = query.Order("id desc")
+	}
+
+	err = query.Find(&users).Error
+	return users, err
 }
 
 func SearchUsers(keyword string) (users []*User, err error) {
@@ -107,8 +123,8 @@ func (user *User) Insert(inviterId int) error {
 		}
 	}
 	user.Quota = config.QuotaForNewUser
-	user.AccessToken = helper.GetUUID()
-	user.AffCode = helper.GetRandomString(4)
+	user.AccessToken = random.GetUUID()
+	user.AffCode = random.GetRandomString(4)
 	result := DB.Create(user)
 	if result.Error != nil {
 		return result.Error
@@ -126,6 +142,22 @@ func (user *User) Insert(inviterId int) error {
 			RecordLog(inviterId, LogTypeSystem, fmt.Sprintf("邀请用户赠送 %s", common.LogQuota(config.QuotaForInviter)))
 		}
 	}
+	// create default token
+	cleanToken := Token{
+		UserId:         user.Id,
+		Name:           "default",
+		Key:            random.GenerateKey(),
+		CreatedTime:    helper.GetTimestamp(),
+		AccessedTime:   helper.GetTimestamp(),
+		ExpiredTime:    -1,
+		RemainQuota:    -1,
+		UnlimitedQuota: true,
+	}
+	result.Error = cleanToken.Insert()
+	if result.Error != nil {
+		// do not block
+		logger.SysError(fmt.Sprintf("create default token for user %d failed: %s", user.Id, result.Error.Error()))
+	}
 	return nil
 }
 
@@ -137,9 +169,9 @@ func (user *User) Update(updatePassword bool) error {
 			return err
 		}
 	}
-	if user.Status == common.UserStatusDisabled {
+	if user.Status == UserStatusDisabled {
 		blacklist.BanUser(user.Id)
-	} else if user.Status == common.UserStatusEnabled {
+	} else if user.Status == UserStatusEnabled {
 		blacklist.UnbanUser(user.Id)
 	}
 	err = DB.Model(user).Updates(user).Error
@@ -151,8 +183,8 @@ func (user *User) Delete() error {
 		return errors.New("id 为空！")
 	}
 	blacklist.BanUser(user.Id)
-	user.Username = fmt.Sprintf("deleted_%s", helper.GetUUID())
-	user.Status = common.UserStatusDeleted
+	user.Username = fmt.Sprintf("deleted_%s", random.GetUUID())
+	user.Status = UserStatusDeleted
 	err := DB.Model(user).Updates(user).Error
 	return err
 }
@@ -176,7 +208,7 @@ func (user *User) ValidateAndFill() (err error) {
 		}
 	}
 	okay := common.ValidatePasswordAndHash(password, user.Password)
-	if !okay || user.Status != common.UserStatusEnabled {
+	if !okay || user.Status != UserStatusEnabled {
 		return errors.New("用户名或密码错误，或用户已被封禁")
 	}
 	return nil
@@ -206,6 +238,22 @@ func (user *User) FillUserByGitHubId() error {
 	return nil
 }
 
+func (user *User) FillUserByLarkId() error {
+	if user.LarkId == "" {
+		return errors.New("lark id 为空！")
+	}
+	DB.Where(User{LarkId: user.LarkId}).First(user)
+	return nil
+}
+
+func (user *User) FillUserByOidcId() error {
+	if user.OidcId == "" {
+		return errors.New("oidc id 为空！")
+	}
+	DB.Where(User{OidcId: user.OidcId}).First(user)
+	return nil
+}
+
 func (user *User) FillUserByWeChatId() error {
 	if user.WeChatId == "" {
 		return errors.New("WeChat id 为空！")
@@ -234,6 +282,14 @@ func IsGitHubIdAlreadyTaken(githubId string) bool {
 	return DB.Where("github_id = ?", githubId).Find(&User{}).RowsAffected == 1
 }
 
+func IsLarkIdAlreadyTaken(githubId string) bool {
+	return DB.Where("lark_id = ?", githubId).Find(&User{}).RowsAffected == 1
+}
+
+func IsOidcIdAlreadyTaken(oidcId string) bool {
+	return DB.Where("oidc_id = ?", oidcId).Find(&User{}).RowsAffected == 1
+}
+
 func IsUsernameAlreadyTaken(username string) bool {
 	return DB.Where("username = ?", username).Find(&User{}).RowsAffected == 1
 }
@@ -260,7 +316,7 @@ func IsAdmin(userId int) bool {
 		logger.SysError("no such user " + err.Error())
 		return false
 	}
-	return user.Role >= common.RoleAdminUser
+	return user.Role >= RoleAdminUser
 }
 
 func IsUserEnabled(userId int) (bool, error) {
@@ -272,7 +328,7 @@ func IsUserEnabled(userId int) (bool, error) {
 	if err != nil {
 		return false, err
 	}
-	return user.Status == common.UserStatusEnabled, nil
+	return user.Status == UserStatusEnabled, nil
 }
 
 func ValidateAccessToken(token string) (user *User) {
@@ -345,7 +401,7 @@ func decreaseUserQuota(id int, quota int64) (err error) {
 }
 
 func GetRootUserEmail() (email string) {
-	DB.Model(&User{}).Where("role = ?", common.RoleRootUser).Select("email").Find(&email)
+	DB.Model(&User{}).Where("role = ?", RoleRootUser).Select("email").Find(&email)
 	return email
 }
 
diff --git a/monitor/channel.go b/monitor/channel.go
index 597ab11a..7e5dc58a 100644
--- a/monitor/channel.go
+++ b/monitor/channel.go
@@ -2,7 +2,6 @@ package monitor
 
 import (
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/common/message"
@@ -29,27 +28,27 @@ func notifyRootUser(subject string, content string) {
 
 // DisableChannel disable & notify
 func DisableChannel(channelId int, channelName string, reason string) {
-	model.UpdateChannelStatusById(channelId, common.ChannelStatusAutoDisabled)
+	model.UpdateChannelStatusById(channelId, model.ChannelStatusAutoDisabled)
 	logger.SysLog(fmt.Sprintf("channel #%d has been disabled: %s", channelId, reason))
-	subject := fmt.Sprintf("通道「%s」（#%d）已被禁用", channelName, channelId)
-	content := fmt.Sprintf("通道「%s」（#%d）已被禁用，原因：%s", channelName, channelId, reason)
+	subject := fmt.Sprintf("渠道「%s」（#%d）已被禁用", channelName, channelId)
+	content := fmt.Sprintf("渠道「%s」（#%d）已被禁用，原因：%s", channelName, channelId, reason)
 	notifyRootUser(subject, content)
 }
 
 func MetricDisableChannel(channelId int, successRate float64) {
-	model.UpdateChannelStatusById(channelId, common.ChannelStatusAutoDisabled)
+	model.UpdateChannelStatusById(channelId, model.ChannelStatusAutoDisabled)
 	logger.SysLog(fmt.Sprintf("channel #%d has been disabled due to low success rate: %.2f", channelId, successRate*100))
-	subject := fmt.Sprintf("通道 #%d 已被禁用", channelId)
-	content := fmt.Sprintf("该渠道在最近 %d 次调用中成功率为 %.2f%%，低于阈值 %.2f%%，因此被系统自动禁用。",
-		config.MetricQueueSize, successRate*100, config.MetricSuccessRateThreshold*100)
+	subject := fmt.Sprintf("渠道 #%d 已被禁用", channelId)
+	content := fmt.Sprintf("该渠道（#%d）在最近 %d 次调用中成功率为 %.2f%%，低于阈值 %.2f%%，因此被系统自动禁用。",
+		channelId, config.MetricQueueSize, successRate*100, config.MetricSuccessRateThreshold*100)
 	notifyRootUser(subject, content)
 }
 
 // EnableChannel enable & notify
 func EnableChannel(channelId int, channelName string) {
-	model.UpdateChannelStatusById(channelId, common.ChannelStatusEnabled)
+	model.UpdateChannelStatusById(channelId, model.ChannelStatusEnabled)
 	logger.SysLog(fmt.Sprintf("channel #%d has been enabled", channelId))
-	subject := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
-	content := fmt.Sprintf("通道「%s」（#%d）已被启用", channelName, channelId)
+	subject := fmt.Sprintf("渠道「%s」（#%d）已被启用", channelName, channelId)
+	content := fmt.Sprintf("渠道「%s」（#%d）已被启用", channelName, channelId)
 	notifyRootUser(subject, content)
 }
diff --git a/monitor/manage.go b/monitor/manage.go
new file mode 100644
index 00000000..268d3924
--- /dev/null
+++ b/monitor/manage.go
@@ -0,0 +1,55 @@
+package monitor
+
+import (
+	"net/http"
+	"strings"
+
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+func ShouldDisableChannel(err *model.Error, statusCode int) bool {
+	if !config.AutomaticDisableChannelEnabled {
+		return false
+	}
+	if err == nil {
+		return false
+	}
+	if statusCode == http.StatusUnauthorized {
+		return true
+	}
+	switch err.Type {
+	case "insufficient_quota", "authentication_error", "permission_error", "forbidden":
+		return true
+	}
+	if err.Code == "invalid_api_key" || err.Code == "account_deactivated" {
+		return true
+	}
+
+	lowerMessage := strings.ToLower(err.Message)
+	if strings.Contains(lowerMessage, "your access was terminated") ||
+		strings.Contains(lowerMessage, "violation of our policies") ||
+		strings.Contains(lowerMessage, "your credit balance is too low") ||
+		strings.Contains(lowerMessage, "organization has been disabled") ||
+		strings.Contains(lowerMessage, "credit") ||
+		strings.Contains(lowerMessage, "balance") ||
+		strings.Contains(lowerMessage, "permission denied") ||
+		strings.Contains(lowerMessage, "organization has been restricted") || // groq
+		strings.Contains(lowerMessage, "已欠费") {
+		return true
+	}
+	return false
+}
+
+func ShouldEnableChannel(err error, openAIErr *model.Error) bool {
+	if !config.AutomaticEnableChannelEnabled {
+		return false
+	}
+	if err != nil {
+		return false
+	}
+	if openAIErr != nil {
+		return false
+	}
+	return true
+}
diff --git a/relay/adaptor.go b/relay/adaptor.go
new file mode 100644
index 00000000..03e83903
--- /dev/null
+++ b/relay/adaptor.go
@@ -0,0 +1,69 @@
+package relay
+
+import (
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/aiproxy"
+	"github.com/songquanpeng/one-api/relay/adaptor/ali"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws"
+	"github.com/songquanpeng/one-api/relay/adaptor/baidu"
+	"github.com/songquanpeng/one-api/relay/adaptor/cloudflare"
+	"github.com/songquanpeng/one-api/relay/adaptor/cohere"
+	"github.com/songquanpeng/one-api/relay/adaptor/coze"
+	"github.com/songquanpeng/one-api/relay/adaptor/deepl"
+	"github.com/songquanpeng/one-api/relay/adaptor/gemini"
+	"github.com/songquanpeng/one-api/relay/adaptor/ollama"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor/palm"
+	"github.com/songquanpeng/one-api/relay/adaptor/proxy"
+	"github.com/songquanpeng/one-api/relay/adaptor/replicate"
+	"github.com/songquanpeng/one-api/relay/adaptor/tencent"
+	"github.com/songquanpeng/one-api/relay/adaptor/vertexai"
+	"github.com/songquanpeng/one-api/relay/adaptor/xunfei"
+	"github.com/songquanpeng/one-api/relay/adaptor/zhipu"
+	"github.com/songquanpeng/one-api/relay/apitype"
+)
+
+func GetAdaptor(apiType int) adaptor.Adaptor {
+	switch apiType {
+	case apitype.AIProxyLibrary:
+		return &aiproxy.Adaptor{}
+	case apitype.Ali:
+		return &ali.Adaptor{}
+	case apitype.Anthropic:
+		return &anthropic.Adaptor{}
+	case apitype.AwsClaude:
+		return &aws.Adaptor{}
+	case apitype.Baidu:
+		return &baidu.Adaptor{}
+	case apitype.Gemini:
+		return &gemini.Adaptor{}
+	case apitype.OpenAI:
+		return &openai.Adaptor{}
+	case apitype.PaLM:
+		return &palm.Adaptor{}
+	case apitype.Tencent:
+		return &tencent.Adaptor{}
+	case apitype.Xunfei:
+		return &xunfei.Adaptor{}
+	case apitype.Zhipu:
+		return &zhipu.Adaptor{}
+	case apitype.Ollama:
+		return &ollama.Adaptor{}
+	case apitype.Coze:
+		return &coze.Adaptor{}
+	case apitype.Cohere:
+		return &cohere.Adaptor{}
+	case apitype.Cloudflare:
+		return &cloudflare.Adaptor{}
+	case apitype.DeepL:
+		return &deepl.Adaptor{}
+	case apitype.VertexAI:
+		return &vertexai.Adaptor{}
+	case apitype.Proxy:
+		return &proxy.Adaptor{}
+	case apitype.Replicate:
+		return &replicate.Adaptor{}
+	}
+	return nil
+}
diff --git a/relay/channel/ai360/constants.go b/relay/adaptor/ai360/constants.go
similarity index 100%
rename from relay/channel/ai360/constants.go
rename to relay/adaptor/ai360/constants.go
diff --git a/relay/channel/aiproxy/adaptor.go b/relay/adaptor/aiproxy/adaptor.go
similarity index 54%
rename from relay/channel/aiproxy/adaptor.go
rename to relay/adaptor/aiproxy/adaptor.go
index 2b4e3022..42d49c0a 100644
--- a/relay/channel/aiproxy/adaptor.go
+++ b/relay/adaptor/aiproxy/adaptor.go
@@ -4,27 +4,27 @@ import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/channel"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 )
 
 type Adaptor struct {
+	meta *meta.Meta
 }
 
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
 }
 
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	return fmt.Sprintf("%s/api/library/ask", meta.BaseURL), nil
 }
 
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
 	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
 	return nil
 }
@@ -34,15 +34,22 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 		return nil, errors.New("request is nil")
 	}
 	aiProxyLibraryRequest := ConvertRequest(*request)
-	aiProxyLibraryRequest.LibraryId = c.GetString(common.ConfigKeyLibraryID)
+	aiProxyLibraryRequest.LibraryId = a.meta.Config.LibraryID
 	return aiProxyLibraryRequest, nil
 }
 
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
 	if meta.IsStream {
 		err, usage = StreamHandler(c, resp)
 	} else {
diff --git a/relay/channel/aiproxy/constants.go b/relay/adaptor/aiproxy/constants.go
similarity index 60%
rename from relay/channel/aiproxy/constants.go
rename to relay/adaptor/aiproxy/constants.go
index c4df51c4..818d2709 100644
--- a/relay/channel/aiproxy/constants.go
+++ b/relay/adaptor/aiproxy/constants.go
@@ -1,6 +1,6 @@
 package aiproxy
 
-import "github.com/songquanpeng/one-api/relay/channel/openai"
+import "github.com/songquanpeng/one-api/relay/adaptor/openai"
 
 var ModelList = []string{""}
 
diff --git a/relay/channel/aiproxy/main.go b/relay/adaptor/aiproxy/main.go
similarity index 73%
rename from relay/channel/aiproxy/main.go
rename to relay/adaptor/aiproxy/main.go
index 96972407..d64b6809 100644
--- a/relay/channel/aiproxy/main.go
+++ b/relay/adaptor/aiproxy/main.go
@@ -4,17 +4,20 @@ import (
 	"bufio"
 	"encoding/json"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strconv"
 	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 // https://docs.aiproxy.io/dev/library#使用已经定制好的知识库进行对话问答
@@ -53,7 +56,7 @@ func responseAIProxyLibrary2OpenAI(response *LibraryResponse) *openai.TextRespon
 		FinishReason: "stop",
 	}
 	fullTextResponse := openai.TextResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
@@ -66,7 +69,7 @@ func documentsAIProxyLibrary(documents []LibraryDocument) *openai.ChatCompletion
 	choice.Delta.Content = aiProxyDocuments2Markdown(documents)
 	choice.FinishReason = &constant.StopFinishReason
 	return &openai.ChatCompletionsStreamResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   "",
@@ -78,7 +81,7 @@ func streamResponseAIProxyLibrary2OpenAI(response *LibraryStreamResponse) *opena
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = response.Content
 	return &openai.ChatCompletionsStreamResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   response.Model,
@@ -88,6 +91,7 @@ func streamResponseAIProxyLibrary2OpenAI(response *LibraryStreamResponse) *opena
 
 func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var usage model.Usage
+	var documents []LibraryDocument
 	scanner := bufio.NewScanner(resp.Body)
 	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
 		if atEOF && len(data) == 0 {
@@ -101,60 +105,48 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		}
 		return 0, nil, nil
 	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 5 { // ignore blank line or wrong format
-				continue
-			}
-			if data[:5] != "data:" {
-				continue
-			}
-			data = data[5:]
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
-	var documents []LibraryDocument
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			var AIProxyLibraryResponse LibraryStreamResponse
-			err := json.Unmarshal([]byte(data), &AIProxyLibraryResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			if len(AIProxyLibraryResponse.Documents) != 0 {
-				documents = AIProxyLibraryResponse.Documents
-			}
-			response := streamResponseAIProxyLibrary2OpenAI(&AIProxyLibraryResponse)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			response := documentsAIProxyLibrary(documents)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 5 || data[:5] != "data:" {
+			continue
 		}
-	})
-	err := resp.Body.Close()
+		data = data[5:]
+
+		var AIProxyLibraryResponse LibraryStreamResponse
+		err := json.Unmarshal([]byte(data), &AIProxyLibraryResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+		if len(AIProxyLibraryResponse.Documents) != 0 {
+			documents = AIProxyLibraryResponse.Documents
+		}
+		response := streamResponseAIProxyLibrary2OpenAI(&AIProxyLibraryResponse)
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	response := documentsAIProxyLibrary(documents)
+	err := render.ObjectData(c, response)
+	if err != nil {
+		logger.SysError(err.Error())
+	}
+	render.Done(c)
+
+	err = resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
+
 	return nil, &usage
 }
 
diff --git a/relay/channel/aiproxy/model.go b/relay/adaptor/aiproxy/model.go
similarity index 100%
rename from relay/channel/aiproxy/model.go
rename to relay/adaptor/aiproxy/model.go
diff --git a/relay/adaptor/ali/adaptor.go b/relay/adaptor/ali/adaptor.go
new file mode 100644
index 00000000..4aa8a11a
--- /dev/null
+++ b/relay/adaptor/ali/adaptor.go
@@ -0,0 +1,105 @@
+package ali
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+	"io"
+	"net/http"
+)
+
+// https://help.aliyun.com/zh/dashscope/developer-reference/api-details
+
+type Adaptor struct {
+	meta *meta.Meta
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	fullRequestURL := ""
+	switch meta.Mode {
+	case relaymode.Embeddings:
+		fullRequestURL = fmt.Sprintf("%s/api/v1/services/embeddings/text-embedding/text-embedding", meta.BaseURL)
+	case relaymode.ImagesGenerations:
+		fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/text2image/image-synthesis", meta.BaseURL)
+	default:
+		fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/text-generation/generation", meta.BaseURL)
+	}
+
+	return fullRequestURL, nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	if meta.IsStream {
+		req.Header.Set("Accept", "text/event-stream")
+		req.Header.Set("X-DashScope-SSE", "enable")
+	}
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+
+	if meta.Mode == relaymode.ImagesGenerations {
+		req.Header.Set("X-DashScope-Async", "enable")
+	}
+	if a.meta.Config.Plugin != "" {
+		req.Header.Set("X-DashScope-Plugin", a.meta.Config.Plugin)
+	}
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	switch relayMode {
+	case relaymode.Embeddings:
+		aliEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return aliEmbeddingRequest, nil
+	default:
+		aliRequest := ConvertRequest(*request)
+		return aliRequest, nil
+	}
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	aliRequest := ConvertImageRequest(*request)
+	return aliRequest, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		switch meta.Mode {
+		case relaymode.Embeddings:
+			err, usage = EmbeddingHandler(c, resp)
+		case relaymode.ImagesGenerations:
+			err, usage = ImageHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "ali"
+}
diff --git a/relay/adaptor/ali/constants.go b/relay/adaptor/ali/constants.go
new file mode 100644
index 00000000..f3d99520
--- /dev/null
+++ b/relay/adaptor/ali/constants.go
@@ -0,0 +1,23 @@
+package ali
+
+var ModelList = []string{
+	"qwen-turbo", "qwen-turbo-latest",
+	"qwen-plus", "qwen-plus-latest",
+	"qwen-max", "qwen-max-latest",
+	"qwen-max-longcontext",
+	"qwen-vl-max", "qwen-vl-max-latest", "qwen-vl-plus", "qwen-vl-plus-latest",
+	"qwen-vl-ocr", "qwen-vl-ocr-latest",
+	"qwen-audio-turbo",
+	"qwen-math-plus", "qwen-math-plus-latest", "qwen-math-turbo", "qwen-math-turbo-latest",
+	"qwen-coder-plus", "qwen-coder-plus-latest", "qwen-coder-turbo", "qwen-coder-turbo-latest",
+	"qwq-32b-preview", "qwen2.5-72b-instruct", "qwen2.5-32b-instruct", "qwen2.5-14b-instruct", "qwen2.5-7b-instruct", "qwen2.5-3b-instruct", "qwen2.5-1.5b-instruct", "qwen2.5-0.5b-instruct",
+	"qwen2-72b-instruct", "qwen2-57b-a14b-instruct", "qwen2-7b-instruct", "qwen2-1.5b-instruct", "qwen2-0.5b-instruct",
+	"qwen1.5-110b-chat", "qwen1.5-72b-chat", "qwen1.5-32b-chat", "qwen1.5-14b-chat", "qwen1.5-7b-chat", "qwen1.5-1.8b-chat", "qwen1.5-0.5b-chat",
+	"qwen-72b-chat", "qwen-14b-chat", "qwen-7b-chat", "qwen-1.8b-chat", "qwen-1.8b-longcontext-chat",
+	"qwen2-vl-7b-instruct", "qwen2-vl-2b-instruct", "qwen-vl-v1", "qwen-vl-chat-v1",
+	"qwen2-audio-instruct", "qwen-audio-chat",
+	"qwen2.5-math-72b-instruct", "qwen2.5-math-7b-instruct", "qwen2.5-math-1.5b-instruct", "qwen2-math-72b-instruct", "qwen2-math-7b-instruct", "qwen2-math-1.5b-instruct",
+	"qwen2.5-coder-32b-instruct", "qwen2.5-coder-14b-instruct", "qwen2.5-coder-7b-instruct", "qwen2.5-coder-3b-instruct", "qwen2.5-coder-1.5b-instruct", "qwen2.5-coder-0.5b-instruct",
+	"text-embedding-v1", "text-embedding-v3", "text-embedding-v2", "text-embedding-async-v2", "text-embedding-async-v1",
+	"ali-stable-diffusion-xl", "ali-stable-diffusion-v1.5", "wanx-v1",
+}
diff --git a/relay/adaptor/ali/image.go b/relay/adaptor/ali/image.go
new file mode 100644
index 00000000..8261803d
--- /dev/null
+++ b/relay/adaptor/ali/image.go
@@ -0,0 +1,192 @@
+package ali
+
+import (
+	"encoding/base64"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+	"strings"
+	"time"
+)
+
+func ImageHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	apiKey := c.Request.Header.Get("Authorization")
+	apiKey = strings.TrimPrefix(apiKey, "Bearer ")
+	responseFormat := c.GetString("response_format")
+
+	var aliTaskResponse TaskResponse
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &aliTaskResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	if aliTaskResponse.Message != "" {
+		logger.SysError("aliAsyncTask err: " + string(responseBody))
+		return openai.ErrorWrapper(errors.New(aliTaskResponse.Message), "ali_async_task_failed", http.StatusInternalServerError), nil
+	}
+
+	aliResponse, _, err := asyncTaskWait(aliTaskResponse.Output.TaskId, apiKey)
+	if err != nil {
+		return openai.ErrorWrapper(err, "ali_async_task_wait_failed", http.StatusInternalServerError), nil
+	}
+
+	if aliResponse.Output.TaskStatus != "SUCCEEDED" {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: aliResponse.Output.Message,
+				Type:    "ali_error",
+				Param:   "",
+				Code:    aliResponse.Output.Code,
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+
+	fullTextResponse := responseAli2OpenAIImage(aliResponse, responseFormat)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, nil
+}
+
+func asyncTask(taskID string, key string) (*TaskResponse, error, []byte) {
+	url := fmt.Sprintf("https://dashscope.aliyuncs.com/api/v1/tasks/%s", taskID)
+
+	var aliResponse TaskResponse
+
+	req, err := http.NewRequest("GET", url, nil)
+	if err != nil {
+		return &aliResponse, err, nil
+	}
+
+	req.Header.Set("Authorization", "Bearer "+key)
+
+	client := &http.Client{}
+	resp, err := client.Do(req)
+	if err != nil {
+		logger.SysError("aliAsyncTask client.Do err: " + err.Error())
+		return &aliResponse, err, nil
+	}
+	defer resp.Body.Close()
+
+	responseBody, err := io.ReadAll(resp.Body)
+
+	var response TaskResponse
+	err = json.Unmarshal(responseBody, &response)
+	if err != nil {
+		logger.SysError("aliAsyncTask NewDecoder err: " + err.Error())
+		return &aliResponse, err, nil
+	}
+
+	return &response, nil, responseBody
+}
+
+func asyncTaskWait(taskID string, key string) (*TaskResponse, []byte, error) {
+	waitSeconds := 2
+	step := 0
+	maxStep := 20
+
+	var taskResponse TaskResponse
+	var responseBody []byte
+
+	for {
+		step++
+		rsp, err, body := asyncTask(taskID, key)
+		responseBody = body
+		if err != nil {
+			return &taskResponse, responseBody, err
+		}
+
+		if rsp.Output.TaskStatus == "" {
+			return &taskResponse, responseBody, nil
+		}
+
+		switch rsp.Output.TaskStatus {
+		case "FAILED":
+			fallthrough
+		case "CANCELED":
+			fallthrough
+		case "SUCCEEDED":
+			fallthrough
+		case "UNKNOWN":
+			return rsp, responseBody, nil
+		}
+		if step >= maxStep {
+			break
+		}
+		time.Sleep(time.Duration(waitSeconds) * time.Second)
+	}
+
+	return nil, nil, fmt.Errorf("aliAsyncTaskWait timeout")
+}
+
+func responseAli2OpenAIImage(response *TaskResponse, responseFormat string) *openai.ImageResponse {
+	imageResponse := openai.ImageResponse{
+		Created: helper.GetTimestamp(),
+	}
+
+	for _, data := range response.Output.Results {
+		var b64Json string
+		if responseFormat == "b64_json" {
+			// 读取 data.Url 的图片数据并转存到 b64Json
+			imageData, err := getImageData(data.Url)
+			if err != nil {
+				// 处理获取图片数据失败的情况
+				logger.SysError("getImageData Error getting image data: " + err.Error())
+				continue
+			}
+
+			// 将图片数据转为 Base64 编码的字符串
+			b64Json = Base64Encode(imageData)
+		} else {
+			// 如果 responseFormat 不是 "b64_json"，则直接使用 data.B64Image
+			b64Json = data.B64Image
+		}
+
+		imageResponse.Data = append(imageResponse.Data, openai.ImageData{
+			Url:           data.Url,
+			B64Json:       b64Json,
+			RevisedPrompt: "",
+		})
+	}
+	return &imageResponse
+}
+
+func getImageData(url string) ([]byte, error) {
+	response, err := http.Get(url)
+	if err != nil {
+		return nil, err
+	}
+	defer response.Body.Close()
+
+	imageData, err := io.ReadAll(response.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	return imageData, nil
+}
+
+func Base64Encode(data []byte) string {
+	b64Json := base64.StdEncoding.EncodeToString(data)
+	return b64Json
+}
diff --git a/relay/channel/ali/main.go b/relay/adaptor/ali/main.go
similarity index 75%
rename from relay/channel/ali/main.go
rename to relay/adaptor/ali/main.go
index 62115d58..6a73c707 100644
--- a/relay/channel/ali/main.go
+++ b/relay/adaptor/ali/main.go
@@ -3,15 +3,18 @@ package ali
 import (
 	"bufio"
 	"encoding/json"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/render"
+	"io"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
-	"strings"
 )
 
 // https://help.aliyun.com/document_detail/613695.html?spm=a2c4g.2399480.0.0.1adb778fAdzP9w#341800c0f8w0r
@@ -33,9 +36,7 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 		enableSearch = true
 		aliModel = strings.TrimSuffix(aliModel, EnableSearchModelSuffix)
 	}
-	if request.TopP >= 1 {
-		request.TopP = 0.9999
-	}
+	request.TopP = helper.Float64PtrMax(request.TopP, 0.9999)
 	return &ChatRequest{
 		Model: aliModel,
 		Input: Input{
@@ -48,13 +49,16 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 			MaxTokens:         request.MaxTokens,
 			Temperature:       request.Temperature,
 			TopP:              request.TopP,
+			TopK:              request.TopK,
+			ResultFormat:      "message",
+			Tools:             request.Tools,
 		},
 	}
 }
 
 func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
 	return &EmbeddingRequest{
-		Model: "text-embedding-v1",
+		Model: request.Model,
 		Input: struct {
 			Texts []string `json:"texts"`
 		}{
@@ -63,6 +67,17 @@ func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingReque
 	}
 }
 
+func ConvertImageRequest(request model.ImageRequest) *ImageRequest {
+	var imageRequest ImageRequest
+	imageRequest.Input.Prompt = request.Prompt
+	imageRequest.Model = request.Model
+	imageRequest.Parameters.Size = strings.Replace(request.Size, "x", "*", -1)
+	imageRequest.Parameters.N = request.N
+	imageRequest.ResponseFormat = request.ResponseFormat
+
+	return &imageRequest
+}
+
 func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var aliResponse EmbeddingResponse
 	err := json.NewDecoder(resp.Body).Decode(&aliResponse)
@@ -86,8 +101,9 @@ func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStat
 			StatusCode: resp.StatusCode,
 		}, nil
 	}
-
+	requestModel := c.GetString(ctxkey.RequestModel)
 	fullTextResponse := embeddingResponseAli2OpenAI(&aliResponse)
+	fullTextResponse.Model = requestModel
 	jsonResponse, err := json.Marshal(fullTextResponse)
 	if err != nil {
 		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
@@ -117,19 +133,11 @@ func embeddingResponseAli2OpenAI(response *EmbeddingResponse) *openai.EmbeddingR
 }
 
 func responseAli2OpenAI(response *ChatResponse) *openai.TextResponse {
-	choice := openai.TextResponseChoice{
-		Index: 0,
-		Message: model.Message{
-			Role:    "assistant",
-			Content: response.Output.Text,
-		},
-		FinishReason: response.Output.FinishReason,
-	}
 	fullTextResponse := openai.TextResponse{
 		Id:      response.RequestId,
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
-		Choices: []openai.TextResponseChoice{choice},
+		Choices: response.Output.Choices,
 		Usage: model.Usage{
 			PromptTokens:     response.Usage.InputTokens,
 			CompletionTokens: response.Usage.OutputTokens,
@@ -140,10 +148,14 @@ func responseAli2OpenAI(response *ChatResponse) *openai.TextResponse {
 }
 
 func streamResponseAli2OpenAI(aliResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
+	if len(aliResponse.Output.Choices) == 0 {
+		return nil
+	}
+	aliChoice := aliResponse.Output.Choices[0]
 	var choice openai.ChatCompletionsStreamResponseChoice
-	choice.Delta.Content = aliResponse.Output.Text
-	if aliResponse.Output.FinishReason != "null" {
-		finishReason := aliResponse.Output.FinishReason
+	choice.Delta = aliChoice.Message
+	if aliChoice.FinishReason != "null" {
+		finishReason := aliChoice.FinishReason
 		choice.FinishReason = &finishReason
 	}
 	response := openai.ChatCompletionsStreamResponse{
@@ -171,53 +183,43 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		}
 		return 0, nil, nil
 	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 5 { // ignore blank line or wrong format
-				continue
-			}
-			if data[:5] != "data:" {
-				continue
-			}
-			data = data[5:]
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
-	//lastResponseText := ""
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			var aliResponse ChatResponse
-			err := json.Unmarshal([]byte(data), &aliResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			if aliResponse.Usage.OutputTokens != 0 {
-				usage.PromptTokens = aliResponse.Usage.InputTokens
-				usage.CompletionTokens = aliResponse.Usage.OutputTokens
-				usage.TotalTokens = aliResponse.Usage.InputTokens + aliResponse.Usage.OutputTokens
-			}
-			response := streamResponseAli2OpenAI(&aliResponse)
-			//response.Choices[0].Delta.Content = strings.TrimPrefix(response.Choices[0].Delta.Content, lastResponseText)
-			//lastResponseText = aliResponse.Output.Text
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 5 || data[:5] != "data:" {
+			continue
 		}
-	})
+		data = data[5:]
+
+		var aliResponse ChatResponse
+		err := json.Unmarshal([]byte(data), &aliResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+		if aliResponse.Usage.OutputTokens != 0 {
+			usage.PromptTokens = aliResponse.Usage.InputTokens
+			usage.CompletionTokens = aliResponse.Usage.OutputTokens
+			usage.TotalTokens = aliResponse.Usage.InputTokens + aliResponse.Usage.OutputTokens
+		}
+		response := streamResponseAli2OpenAI(&aliResponse)
+		if response == nil {
+			continue
+		}
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
@@ -226,6 +228,7 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 }
 
 func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	ctx := c.Request.Context()
 	var aliResponse ChatResponse
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
@@ -235,6 +238,7 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
+	logger.Debugf(ctx, "response body: %s\n", responseBody)
 	err = json.Unmarshal(responseBody, &aliResponse)
 	if err != nil {
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
diff --git a/relay/adaptor/ali/model.go b/relay/adaptor/ali/model.go
new file mode 100644
index 00000000..a680c7e2
--- /dev/null
+++ b/relay/adaptor/ali/model.go
@@ -0,0 +1,154 @@
+package ali
+
+import (
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type Message struct {
+	Content string `json:"content"`
+	Role    string `json:"role"`
+}
+
+type Input struct {
+	//Prompt   string       `json:"prompt"`
+	Messages []Message `json:"messages"`
+}
+
+type Parameters struct {
+	TopP              *float64     `json:"top_p,omitempty"`
+	TopK              int          `json:"top_k,omitempty"`
+	Seed              uint64       `json:"seed,omitempty"`
+	EnableSearch      bool         `json:"enable_search,omitempty"`
+	IncrementalOutput bool         `json:"incremental_output,omitempty"`
+	MaxTokens         int          `json:"max_tokens,omitempty"`
+	Temperature       *float64     `json:"temperature,omitempty"`
+	ResultFormat      string       `json:"result_format,omitempty"`
+	Tools             []model.Tool `json:"tools,omitempty"`
+}
+
+type ChatRequest struct {
+	Model      string     `json:"model"`
+	Input      Input      `json:"input"`
+	Parameters Parameters `json:"parameters,omitempty"`
+}
+
+type ImageRequest struct {
+	Model string `json:"model"`
+	Input struct {
+		Prompt         string `json:"prompt"`
+		NegativePrompt string `json:"negative_prompt,omitempty"`
+	} `json:"input"`
+	Parameters struct {
+		Size  string `json:"size,omitempty"`
+		N     int    `json:"n,omitempty"`
+		Steps string `json:"steps,omitempty"`
+		Scale string `json:"scale,omitempty"`
+	} `json:"parameters,omitempty"`
+	ResponseFormat string `json:"response_format,omitempty"`
+}
+
+type TaskResponse struct {
+	StatusCode int    `json:"status_code,omitempty"`
+	RequestId  string `json:"request_id,omitempty"`
+	Code       string `json:"code,omitempty"`
+	Message    string `json:"message,omitempty"`
+	Output     struct {
+		TaskId     string `json:"task_id,omitempty"`
+		TaskStatus string `json:"task_status,omitempty"`
+		Code       string `json:"code,omitempty"`
+		Message    string `json:"message,omitempty"`
+		Results    []struct {
+			B64Image string `json:"b64_image,omitempty"`
+			Url      string `json:"url,omitempty"`
+			Code     string `json:"code,omitempty"`
+			Message  string `json:"message,omitempty"`
+		} `json:"results,omitempty"`
+		TaskMetrics struct {
+			Total     int `json:"TOTAL,omitempty"`
+			Succeeded int `json:"SUCCEEDED,omitempty"`
+			Failed    int `json:"FAILED,omitempty"`
+		} `json:"task_metrics,omitempty"`
+	} `json:"output,omitempty"`
+	Usage Usage `json:"usage"`
+}
+
+type Header struct {
+	Action       string `json:"action,omitempty"`
+	Streaming    string `json:"streaming,omitempty"`
+	TaskID       string `json:"task_id,omitempty"`
+	Event        string `json:"event,omitempty"`
+	ErrorCode    string `json:"error_code,omitempty"`
+	ErrorMessage string `json:"error_message,omitempty"`
+	Attributes   any    `json:"attributes,omitempty"`
+}
+
+type Payload struct {
+	Model      string `json:"model,omitempty"`
+	Task       string `json:"task,omitempty"`
+	TaskGroup  string `json:"task_group,omitempty"`
+	Function   string `json:"function,omitempty"`
+	Parameters struct {
+		SampleRate int     `json:"sample_rate,omitempty"`
+		Rate       float64 `json:"rate,omitempty"`
+		Format     string  `json:"format,omitempty"`
+	} `json:"parameters,omitempty"`
+	Input struct {
+		Text string `json:"text,omitempty"`
+	} `json:"input,omitempty"`
+	Usage struct {
+		Characters int `json:"characters,omitempty"`
+	} `json:"usage,omitempty"`
+}
+
+type WSSMessage struct {
+	Header  Header  `json:"header,omitempty"`
+	Payload Payload `json:"payload,omitempty"`
+}
+
+type EmbeddingRequest struct {
+	Model string `json:"model"`
+	Input struct {
+		Texts []string `json:"texts"`
+	} `json:"input"`
+	Parameters *struct {
+		TextType string `json:"text_type,omitempty"`
+	} `json:"parameters,omitempty"`
+}
+
+type Embedding struct {
+	Embedding []float64 `json:"embedding"`
+	TextIndex int       `json:"text_index"`
+}
+
+type EmbeddingResponse struct {
+	Output struct {
+		Embeddings []Embedding `json:"embeddings"`
+	} `json:"output"`
+	Usage Usage `json:"usage"`
+	Error
+}
+
+type Error struct {
+	Code      string `json:"code"`
+	Message   string `json:"message"`
+	RequestId string `json:"request_id"`
+}
+
+type Usage struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
+	TotalTokens  int `json:"total_tokens"`
+}
+
+type Output struct {
+	//Text         string                      `json:"text"`
+	//FinishReason string                      `json:"finish_reason"`
+	Choices []openai.TextResponseChoice `json:"choices"`
+}
+
+type ChatResponse struct {
+	Output Output `json:"output"`
+	Usage  Usage  `json:"usage"`
+	Error
+}
diff --git a/relay/channel/anthropic/adaptor.go b/relay/adaptor/anthropic/adaptor.go
similarity index 54%
rename from relay/channel/anthropic/adaptor.go
rename to relay/adaptor/anthropic/adaptor.go
index a165b35c..bd0949be 100644
--- a/relay/channel/anthropic/adaptor.go
+++ b/relay/adaptor/anthropic/adaptor.go
@@ -3,27 +3,29 @@ package anthropic
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 type Adaptor struct {
 }
 
-func (a *Adaptor) Init(meta *util.RelayMeta) {
+func (a *Adaptor) Init(meta *meta.Meta) {
 
 }
 
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	return fmt.Sprintf("%s/v1/messages", meta.BaseURL), nil
 }
 
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
 	req.Header.Set("x-api-key", meta.APIKey)
 	anthropicVersion := c.Request.Header.Get("anthropic-version")
 	if anthropicVersion == "" {
@@ -31,6 +33,13 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *ut
 	}
 	req.Header.Set("anthropic-version", anthropicVersion)
 	req.Header.Set("anthropic-beta", "messages-2023-12-15")
+
+	// https://x.com/alexalbert__/status/1812921642143900036
+	// claude-3-5-sonnet can support 8k context
+	if strings.HasPrefix(meta.ActualModelName, "claude-3-5-sonnet") {
+		req.Header.Set("anthropic-beta", "max-tokens-3-5-sonnet-2024-07-15")
+	}
+
 	return nil
 }
 
@@ -41,11 +50,18 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	return ConvertRequest(*request), nil
 }
 
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
 	if meta.IsStream {
 		err, usage = StreamHandler(c, resp)
 	} else {
diff --git a/relay/channel/anthropic/constants.go b/relay/adaptor/anthropic/constants.go
similarity index 60%
rename from relay/channel/anthropic/constants.go
rename to relay/adaptor/anthropic/constants.go
index cadcedc8..8ea7c4d8 100644
--- a/relay/channel/anthropic/constants.go
+++ b/relay/adaptor/anthropic/constants.go
@@ -3,6 +3,10 @@ package anthropic
 var ModelList = []string{
 	"claude-instant-1.2", "claude-2.0", "claude-2.1",
 	"claude-3-haiku-20240307",
+	"claude-3-5-haiku-20241022",
 	"claude-3-sonnet-20240229",
 	"claude-3-opus-20240229",
+	"claude-3-5-sonnet-20240620",
+	"claude-3-5-sonnet-20241022",
+	"claude-3-5-sonnet-latest",
 }
diff --git a/relay/channel/anthropic/main.go b/relay/adaptor/anthropic/main.go
similarity index 55%
rename from relay/channel/anthropic/main.go
rename to relay/adaptor/anthropic/main.go
index 3eeb0b2c..d3e306c8 100644
--- a/relay/channel/anthropic/main.go
+++ b/relay/adaptor/anthropic/main.go
@@ -4,16 +4,18 @@ import (
 	"bufio"
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common/render"
+	"io"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/image"
 	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
-	"strings"
 )
 
 func stopReasonClaude2OpenAI(reason *string) string {
@@ -27,18 +29,55 @@ func stopReasonClaude2OpenAI(reason *string) string {
 		return "stop"
 	case "max_tokens":
 		return "length"
+	case "tool_use":
+		return "tool_calls"
 	default:
 		return *reason
 	}
 }
 
 func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
+	claudeTools := make([]Tool, 0, len(textRequest.Tools))
+
+	for _, tool := range textRequest.Tools {
+		if params, ok := tool.Function.Parameters.(map[string]any); ok {
+			claudeTools = append(claudeTools, Tool{
+				Name:        tool.Function.Name,
+				Description: tool.Function.Description,
+				InputSchema: InputSchema{
+					Type:       params["type"].(string),
+					Properties: params["properties"],
+					Required:   params["required"],
+				},
+			})
+		}
+	}
+
 	claudeRequest := Request{
 		Model:       textRequest.Model,
 		MaxTokens:   textRequest.MaxTokens,
 		Temperature: textRequest.Temperature,
 		TopP:        textRequest.TopP,
+		TopK:        textRequest.TopK,
 		Stream:      textRequest.Stream,
+		Tools:       claudeTools,
+	}
+	if len(claudeTools) > 0 {
+		claudeToolChoice := struct {
+			Type string `json:"type"`
+			Name string `json:"name,omitempty"`
+		}{Type: "auto"} // default value https://docs.anthropic.com/en/docs/build-with-claude/tool-use#controlling-claudes-output
+		if choice, ok := textRequest.ToolChoice.(map[string]any); ok {
+			if function, ok := choice["function"].(map[string]any); ok {
+				claudeToolChoice.Type = "tool"
+				claudeToolChoice.Name = function["name"].(string)
+			}
+		} else if toolChoiceType, ok := textRequest.ToolChoice.(string); ok {
+			if toolChoiceType == "any" {
+				claudeToolChoice.Type = toolChoiceType
+			}
+		}
+		claudeRequest.ToolChoice = claudeToolChoice
 	}
 	if claudeRequest.MaxTokens == 0 {
 		claudeRequest.MaxTokens = 4096
@@ -61,7 +100,24 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 		if message.IsStringContent() {
 			content.Type = "text"
 			content.Text = message.StringContent()
+			if message.Role == "tool" {
+				claudeMessage.Role = "user"
+				content.Type = "tool_result"
+				content.Content = content.Text
+				content.Text = ""
+				content.ToolUseId = message.ToolCallId
+			}
 			claudeMessage.Content = append(claudeMessage.Content, content)
+			for i := range message.ToolCalls {
+				inputParam := make(map[string]any)
+				_ = json.Unmarshal([]byte(message.ToolCalls[i].Function.Arguments.(string)), &inputParam)
+				claudeMessage.Content = append(claudeMessage.Content, Content{
+					Type:  "tool_use",
+					Id:    message.ToolCalls[i].Id,
+					Name:  message.ToolCalls[i].Function.Name,
+					Input: inputParam,
+				})
+			}
 			claudeRequest.Messages = append(claudeRequest.Messages, claudeMessage)
 			continue
 		}
@@ -90,20 +146,39 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
 }
 
 // https://docs.anthropic.com/claude/reference/messages-streaming
-func streamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCompletionsStreamResponse, *Response) {
+func StreamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCompletionsStreamResponse, *Response) {
 	var response *Response
 	var responseText string
 	var stopReason string
+	tools := make([]model.Tool, 0)
+
 	switch claudeResponse.Type {
 	case "message_start":
 		return nil, claudeResponse.Message
 	case "content_block_start":
 		if claudeResponse.ContentBlock != nil {
 			responseText = claudeResponse.ContentBlock.Text
+			if claudeResponse.ContentBlock.Type == "tool_use" {
+				tools = append(tools, model.Tool{
+					Id:   claudeResponse.ContentBlock.Id,
+					Type: "function",
+					Function: model.Function{
+						Name:      claudeResponse.ContentBlock.Name,
+						Arguments: "",
+					},
+				})
+			}
 		}
 	case "content_block_delta":
 		if claudeResponse.Delta != nil {
 			responseText = claudeResponse.Delta.Text
+			if claudeResponse.Delta.Type == "input_json_delta" {
+				tools = append(tools, model.Tool{
+					Function: model.Function{
+						Arguments: claudeResponse.Delta.PartialJson,
+					},
+				})
+			}
 		}
 	case "message_delta":
 		if claudeResponse.Usage != nil {
@@ -117,6 +192,10 @@ func streamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCo
 	}
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = responseText
+	if len(tools) > 0 {
+		choice.Delta.Content = nil // compatible with other OpenAI derivative applications, like LobeOpenAICompatibleFactory ...
+		choice.Delta.ToolCalls = tools
+	}
 	choice.Delta.Role = "assistant"
 	finishReason := stopReasonClaude2OpenAI(&stopReason)
 	if finishReason != "null" {
@@ -128,17 +207,32 @@ func streamResponseClaude2OpenAI(claudeResponse *StreamResponse) (*openai.ChatCo
 	return &openaiResponse, response
 }
 
-func responseClaude2OpenAI(claudeResponse *Response) *openai.TextResponse {
+func ResponseClaude2OpenAI(claudeResponse *Response) *openai.TextResponse {
 	var responseText string
 	if len(claudeResponse.Content) > 0 {
 		responseText = claudeResponse.Content[0].Text
 	}
+	tools := make([]model.Tool, 0)
+	for _, v := range claudeResponse.Content {
+		if v.Type == "tool_use" {
+			args, _ := json.Marshal(v.Input)
+			tools = append(tools, model.Tool{
+				Id:   v.Id,
+				Type: "function", // compatible with other OpenAI derivative applications
+				Function: model.Function{
+					Name:      v.Name,
+					Arguments: string(args),
+				},
+			})
+		}
+	}
 	choice := openai.TextResponseChoice{
 		Index: 0,
 		Message: model.Message{
-			Role:    "assistant",
-			Content: responseText,
-			Name:    nil,
+			Role:      "assistant",
+			Content:   responseText,
+			Name:      nil,
+			ToolCalls: tools,
 		},
 		FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
 	}
@@ -167,64 +261,77 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		}
 		return 0, nil, nil
 	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 6 {
-				continue
-			}
-			if !strings.HasPrefix(data, "data: ") {
-				continue
-			}
-			data = strings.TrimPrefix(data, "data: ")
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
+
 	var usage model.Usage
 	var modelName string
 	var id string
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			// some implementations may add \r at the end of data
-			data = strings.TrimSuffix(data, "\r")
-			var claudeResponse StreamResponse
-			err := json.Unmarshal([]byte(data), &claudeResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			response, meta := streamResponseClaude2OpenAI(&claudeResponse)
-			if meta != nil {
-				usage.PromptTokens += meta.Usage.InputTokens
-				usage.CompletionTokens += meta.Usage.OutputTokens
+	var lastToolCallChoice openai.ChatCompletionsStreamResponseChoice
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 6 || !strings.HasPrefix(data, "data:") {
+			continue
+		}
+		data = strings.TrimPrefix(data, "data:")
+		data = strings.TrimSpace(data)
+
+		var claudeResponse StreamResponse
+		err := json.Unmarshal([]byte(data), &claudeResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		response, meta := StreamResponseClaude2OpenAI(&claudeResponse)
+		if meta != nil {
+			usage.PromptTokens += meta.Usage.InputTokens
+			usage.CompletionTokens += meta.Usage.OutputTokens
+			if len(meta.Id) > 0 { // only message_start has an id, otherwise it's a finish_reason event.
 				modelName = meta.Model
 				id = fmt.Sprintf("chatcmpl-%s", meta.Id)
-				return true
+				continue
+			} else { // finish_reason case
+				if len(lastToolCallChoice.Delta.ToolCalls) > 0 {
+					lastArgs := &lastToolCallChoice.Delta.ToolCalls[len(lastToolCallChoice.Delta.ToolCalls)-1].Function
+					if len(lastArgs.Arguments.(string)) == 0 { // compatible with OpenAI sending an empty object `{}` when no arguments.
+						lastArgs.Arguments = "{}"
+						response.Choices[len(response.Choices)-1].Delta.Content = nil
+						response.Choices[len(response.Choices)-1].Delta.ToolCalls = lastToolCallChoice.Delta.ToolCalls
+					}
+				}
 			}
-			if response == nil {
-				return true
-			}
-			response.Id = id
-			response.Model = modelName
-			response.Created = createdTime
-			jsonStr, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
 		}
-	})
-	_ = resp.Body.Close()
+		if response == nil {
+			continue
+		}
+
+		response.Id = id
+		response.Model = modelName
+		response.Created = createdTime
+
+		for _, choice := range response.Choices {
+			if len(choice.Delta.ToolCalls) > 0 {
+				lastToolCallChoice = choice
+			}
+		}
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
+	err := resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
 	return nil, &usage
 }
 
@@ -253,7 +360,7 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName st
 			StatusCode: resp.StatusCode,
 		}, nil
 	}
-	fullTextResponse := responseClaude2OpenAI(&claudeResponse)
+	fullTextResponse := ResponseClaude2OpenAI(&claudeResponse)
 	fullTextResponse.Model = modelName
 	usage := model.Usage{
 		PromptTokens:     claudeResponse.Usage.InputTokens,
diff --git a/relay/channel/anthropic/model.go b/relay/adaptor/anthropic/model.go
similarity index 70%
rename from relay/channel/anthropic/model.go
rename to relay/adaptor/anthropic/model.go
index 32b187cd..47f193fa 100644
--- a/relay/channel/anthropic/model.go
+++ b/relay/adaptor/anthropic/model.go
@@ -16,6 +16,12 @@ type Content struct {
 	Type   string       `json:"type"`
 	Text   string       `json:"text,omitempty"`
 	Source *ImageSource `json:"source,omitempty"`
+	// tool_calls
+	Id        string `json:"id,omitempty"`
+	Name      string `json:"name,omitempty"`
+	Input     any    `json:"input,omitempty"`
+	Content   string `json:"content,omitempty"`
+	ToolUseId string `json:"tool_use_id,omitempty"`
 }
 
 type Message struct {
@@ -23,6 +29,18 @@ type Message struct {
 	Content []Content `json:"content"`
 }
 
+type Tool struct {
+	Name        string      `json:"name"`
+	Description string      `json:"description,omitempty"`
+	InputSchema InputSchema `json:"input_schema"`
+}
+
+type InputSchema struct {
+	Type       string `json:"type"`
+	Properties any    `json:"properties,omitempty"`
+	Required   any    `json:"required,omitempty"`
+}
+
 type Request struct {
 	Model         string    `json:"model"`
 	Messages      []Message `json:"messages"`
@@ -30,9 +48,11 @@ type Request struct {
 	MaxTokens     int       `json:"max_tokens,omitempty"`
 	StopSequences []string  `json:"stop_sequences,omitempty"`
 	Stream        bool      `json:"stream,omitempty"`
-	Temperature   float64   `json:"temperature,omitempty"`
-	TopP          float64   `json:"top_p,omitempty"`
+	Temperature   *float64  `json:"temperature,omitempty"`
+	TopP          *float64  `json:"top_p,omitempty"`
 	TopK          int       `json:"top_k,omitempty"`
+	Tools         []Tool    `json:"tools,omitempty"`
+	ToolChoice    any       `json:"tool_choice,omitempty"`
 	//Metadata    `json:"metadata,omitempty"`
 }
 
@@ -61,6 +81,7 @@ type Response struct {
 type Delta struct {
 	Type         string  `json:"type"`
 	Text         string  `json:"text"`
+	PartialJson  string  `json:"partial_json,omitempty"`
 	StopReason   *string `json:"stop_reason"`
 	StopSequence *string `json:"stop_sequence"`
 }
diff --git a/relay/adaptor/aws/adaptor.go b/relay/adaptor/aws/adaptor.go
new file mode 100644
index 00000000..62221346
--- /dev/null
+++ b/relay/adaptor/aws/adaptor.go
@@ -0,0 +1,84 @@
+package aws
+
+import (
+	"errors"
+	"io"
+	"net/http"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/credentials"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ adaptor.Adaptor = new(Adaptor)
+
+type Adaptor struct {
+	awsAdapter utils.AwsAdapter
+
+	Meta      *meta.Meta
+	AwsClient *bedrockruntime.Client
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.Meta = meta
+	a.AwsClient = bedrockruntime.New(bedrockruntime.Options{
+		Region:      meta.Config.Region,
+		Credentials: aws.NewCredentialsCache(credentials.NewStaticCredentialsProvider(meta.Config.AK, meta.Config.SK, "")),
+	})
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	adaptor := GetAdaptor(request.Model)
+	if adaptor == nil {
+		return nil, errors.New("adaptor not found")
+	}
+
+	a.awsAdapter = adaptor
+	return adaptor.ConvertRequest(c, relayMode, request)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if a.awsAdapter == nil {
+		return nil, utils.WrapErr(errors.New("awsAdapter is nil"))
+	}
+	return a.awsAdapter.DoResponse(c, a.AwsClient, meta)
+}
+
+func (a *Adaptor) GetModelList() (models []string) {
+	for model := range adaptors {
+		models = append(models, model)
+	}
+	return
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "aws"
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return "", nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return nil, nil
+}
diff --git a/relay/adaptor/aws/claude/adapter.go b/relay/adaptor/aws/claude/adapter.go
new file mode 100644
index 00000000..eb3c9fb8
--- /dev/null
+++ b/relay/adaptor/aws/claude/adapter.go
@@ -0,0 +1,37 @@
+package aws
+
+import (
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ utils.AwsAdapter = new(Adaptor)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	claudeReq := anthropic.ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, claudeReq)
+	return claudeReq, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, awsCli)
+	} else {
+		err, usage = Handler(c, awsCli, meta.ActualModelName)
+	}
+	return
+}
diff --git a/relay/adaptor/aws/claude/main.go b/relay/adaptor/aws/claude/main.go
new file mode 100644
index 00000000..3fe3dfd8
--- /dev/null
+++ b/relay/adaptor/aws/claude/main.go
@@ -0,0 +1,207 @@
+// Package aws provides the AWS adaptor for the relay service.
+package aws
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime/types"
+	"github.com/gin-gonic/gin"
+	"github.com/jinzhu/copier"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html
+var AwsModelIDMap = map[string]string{
+	"claude-instant-1.2":         "anthropic.claude-instant-v1",
+	"claude-2.0":                 "anthropic.claude-v2",
+	"claude-2.1":                 "anthropic.claude-v2:1",
+	"claude-3-haiku-20240307":    "anthropic.claude-3-haiku-20240307-v1:0",
+	"claude-3-sonnet-20240229":   "anthropic.claude-3-sonnet-20240229-v1:0",
+	"claude-3-opus-20240229":     "anthropic.claude-3-opus-20240229-v1:0",
+	"claude-3-5-sonnet-20240620": "anthropic.claude-3-5-sonnet-20240620-v1:0",
+	"claude-3-5-sonnet-20241022": "anthropic.claude-3-5-sonnet-20241022-v2:0",
+	"claude-3-5-sonnet-latest":   "anthropic.claude-3-5-sonnet-20241022-v2:0",
+	"claude-3-5-haiku-20241022":  "anthropic.claude-3-5-haiku-20241022-v1:0",
+}
+
+func awsModelID(requestModel string) (string, error) {
+	if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
+		return awsModelID, nil
+	}
+
+	return "", errors.Errorf("model %s not found", requestModel)
+}
+
+func Handler(c *gin.Context, awsCli *bedrockruntime.Client, modelName string) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	claudeReq_, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+	claudeReq := claudeReq_.(*anthropic.Request)
+	awsClaudeReq := &Request{
+		AnthropicVersion: "bedrock-2023-05-31",
+	}
+	if err = copier.Copy(awsClaudeReq, claudeReq); err != nil {
+		return utils.WrapErr(errors.Wrap(err, "copy request")), nil
+	}
+
+	awsReq.Body, err = json.Marshal(awsClaudeReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModel(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModel")), nil
+	}
+
+	claudeResponse := new(anthropic.Response)
+	err = json.Unmarshal(awsResp.Body, claudeResponse)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "unmarshal response")), nil
+	}
+
+	openaiResp := anthropic.ResponseClaude2OpenAI(claudeResponse)
+	openaiResp.Model = modelName
+	usage := relaymodel.Usage{
+		PromptTokens:     claudeResponse.Usage.InputTokens,
+		CompletionTokens: claudeResponse.Usage.OutputTokens,
+		TotalTokens:      claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens,
+	}
+	openaiResp.Usage = usage
+
+	c.JSON(http.StatusOK, openaiResp)
+	return nil, &usage
+}
+
+func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	createdTime := helper.GetTimestamp()
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelWithResponseStreamInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	claudeReq_, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+	claudeReq := claudeReq_.(*anthropic.Request)
+
+	awsClaudeReq := &Request{
+		AnthropicVersion: "bedrock-2023-05-31",
+	}
+	if err = copier.Copy(awsClaudeReq, claudeReq); err != nil {
+		return utils.WrapErr(errors.Wrap(err, "copy request")), nil
+	}
+	awsReq.Body, err = json.Marshal(awsClaudeReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModelWithResponseStream(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModelWithResponseStream")), nil
+	}
+	stream := awsResp.GetStream()
+	defer stream.Close()
+
+	c.Writer.Header().Set("Content-Type", "text/event-stream")
+	var usage relaymodel.Usage
+	var id string
+	var lastToolCallChoice openai.ChatCompletionsStreamResponseChoice
+
+	c.Stream(func(w io.Writer) bool {
+		event, ok := <-stream.Events()
+		if !ok {
+			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
+			return false
+		}
+
+		switch v := event.(type) {
+		case *types.ResponseStreamMemberChunk:
+			claudeResp := new(anthropic.StreamResponse)
+			err := json.NewDecoder(bytes.NewReader(v.Value.Bytes)).Decode(claudeResp)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				return false
+			}
+
+			response, meta := anthropic.StreamResponseClaude2OpenAI(claudeResp)
+			if meta != nil {
+				usage.PromptTokens += meta.Usage.InputTokens
+				usage.CompletionTokens += meta.Usage.OutputTokens
+				if len(meta.Id) > 0 { // only message_start has an id, otherwise it's a finish_reason event.
+					id = fmt.Sprintf("chatcmpl-%s", meta.Id)
+					return true
+				} else { // finish_reason case
+					if len(lastToolCallChoice.Delta.ToolCalls) > 0 {
+						lastArgs := &lastToolCallChoice.Delta.ToolCalls[len(lastToolCallChoice.Delta.ToolCalls)-1].Function
+						if len(lastArgs.Arguments.(string)) == 0 { // compatible with OpenAI sending an empty object `{}` when no arguments.
+							lastArgs.Arguments = "{}"
+							response.Choices[len(response.Choices)-1].Delta.Content = nil
+							response.Choices[len(response.Choices)-1].Delta.ToolCalls = lastToolCallChoice.Delta.ToolCalls
+						}
+					}
+				}
+			}
+			if response == nil {
+				return true
+			}
+			response.Id = id
+			response.Model = c.GetString(ctxkey.OriginalModel)
+			response.Created = createdTime
+
+			for _, choice := range response.Choices {
+				if len(choice.Delta.ToolCalls) > 0 {
+					lastToolCallChoice = choice
+				}
+			}
+			jsonStr, err := json.Marshal(response)
+			if err != nil {
+				logger.SysError("error marshalling stream response: " + err.Error())
+				return true
+			}
+			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
+			return true
+		case *types.UnknownUnionMember:
+			fmt.Println("unknown tag:", v.Tag)
+			return false
+		default:
+			fmt.Println("union is nil or unknown type")
+			return false
+		}
+	})
+
+	return nil, &usage
+}
diff --git a/relay/adaptor/aws/claude/model.go b/relay/adaptor/aws/claude/model.go
new file mode 100644
index 00000000..10622887
--- /dev/null
+++ b/relay/adaptor/aws/claude/model.go
@@ -0,0 +1,20 @@
+package aws
+
+import "github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+
+// Request is the request to AWS Claude
+//
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-anthropic-claude-messages.html
+type Request struct {
+	// AnthropicVersion should be "bedrock-2023-05-31"
+	AnthropicVersion string              `json:"anthropic_version"`
+	Messages         []anthropic.Message `json:"messages"`
+	System           string              `json:"system,omitempty"`
+	MaxTokens        int                 `json:"max_tokens,omitempty"`
+	Temperature      *float64            `json:"temperature,omitempty"`
+	TopP             *float64            `json:"top_p,omitempty"`
+	TopK             int                 `json:"top_k,omitempty"`
+	StopSequences    []string            `json:"stop_sequences,omitempty"`
+	Tools            []anthropic.Tool    `json:"tools,omitempty"`
+	ToolChoice       any                 `json:"tool_choice,omitempty"`
+}
diff --git a/relay/adaptor/aws/llama3/adapter.go b/relay/adaptor/aws/llama3/adapter.go
new file mode 100644
index 00000000..83edbc9d
--- /dev/null
+++ b/relay/adaptor/aws/llama3/adapter.go
@@ -0,0 +1,37 @@
+package aws
+
+import (
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ utils.AwsAdapter = new(Adaptor)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	llamaReq := ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, llamaReq)
+	return llamaReq, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, awsCli)
+	} else {
+		err, usage = Handler(c, awsCli, meta.ActualModelName)
+	}
+	return
+}
diff --git a/relay/adaptor/aws/llama3/main.go b/relay/adaptor/aws/llama3/main.go
new file mode 100644
index 00000000..e5fcd89f
--- /dev/null
+++ b/relay/adaptor/aws/llama3/main.go
@@ -0,0 +1,231 @@
+// Package aws provides the AWS adaptor for the relay service.
+package aws
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"text/template"
+
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/random"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime/types"
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+// Only support llama-3-8b and llama-3-70b instruction models
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html
+var AwsModelIDMap = map[string]string{
+	"llama3-8b-8192":  "meta.llama3-8b-instruct-v1:0",
+	"llama3-70b-8192": "meta.llama3-70b-instruct-v1:0",
+}
+
+func awsModelID(requestModel string) (string, error) {
+	if awsModelID, ok := AwsModelIDMap[requestModel]; ok {
+		return awsModelID, nil
+	}
+
+	return "", errors.Errorf("model %s not found", requestModel)
+}
+
+// promptTemplate with range
+const promptTemplate = `<|begin_of_text|>{{range .Messages}}<|start_header_id|>{{.Role}}<|end_header_id|>{{.StringContent}}<|eot_id|>{{end}}<|start_header_id|>assistant<|end_header_id|>
+`
+
+var promptTpl = template.Must(template.New("llama3-chat").Parse(promptTemplate))
+
+func RenderPrompt(messages []relaymodel.Message) string {
+	var buf bytes.Buffer
+	err := promptTpl.Execute(&buf, struct{ Messages []relaymodel.Message }{messages})
+	if err != nil {
+		logger.SysError("error rendering prompt messages: " + err.Error())
+	}
+	return buf.String()
+}
+
+func ConvertRequest(textRequest relaymodel.GeneralOpenAIRequest) *Request {
+	llamaRequest := Request{
+		MaxGenLen:   textRequest.MaxTokens,
+		Temperature: textRequest.Temperature,
+		TopP:        textRequest.TopP,
+	}
+	if llamaRequest.MaxGenLen == 0 {
+		llamaRequest.MaxGenLen = 2048
+	}
+	prompt := RenderPrompt(textRequest.Messages)
+	llamaRequest.Prompt = prompt
+	return &llamaRequest
+}
+
+func Handler(c *gin.Context, awsCli *bedrockruntime.Client, modelName string) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	llamaReq, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+
+	awsReq.Body, err = json.Marshal(llamaReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModel(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModel")), nil
+	}
+
+	var llamaResponse Response
+	err = json.Unmarshal(awsResp.Body, &llamaResponse)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "unmarshal response")), nil
+	}
+
+	openaiResp := ResponseLlama2OpenAI(&llamaResponse)
+	openaiResp.Model = modelName
+	usage := relaymodel.Usage{
+		PromptTokens:     llamaResponse.PromptTokenCount,
+		CompletionTokens: llamaResponse.GenerationTokenCount,
+		TotalTokens:      llamaResponse.PromptTokenCount + llamaResponse.GenerationTokenCount,
+	}
+	openaiResp.Usage = usage
+
+	c.JSON(http.StatusOK, openaiResp)
+	return nil, &usage
+}
+
+func ResponseLlama2OpenAI(llamaResponse *Response) *openai.TextResponse {
+	var responseText string
+	if len(llamaResponse.Generation) > 0 {
+		responseText = llamaResponse.Generation
+	}
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: relaymodel.Message{
+			Role:    "assistant",
+			Content: responseText,
+			Name:    nil,
+		},
+		FinishReason: llamaResponse.StopReason,
+	}
+	fullTextResponse := openai.TextResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
+		Object:  "chat.completion",
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+	}
+	return &fullTextResponse
+}
+
+func StreamHandler(c *gin.Context, awsCli *bedrockruntime.Client) (*relaymodel.ErrorWithStatusCode, *relaymodel.Usage) {
+	createdTime := helper.GetTimestamp()
+	awsModelId, err := awsModelID(c.GetString(ctxkey.RequestModel))
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "awsModelID")), nil
+	}
+
+	awsReq := &bedrockruntime.InvokeModelWithResponseStreamInput{
+		ModelId:     aws.String(awsModelId),
+		Accept:      aws.String("application/json"),
+		ContentType: aws.String("application/json"),
+	}
+
+	llamaReq, ok := c.Get(ctxkey.ConvertedRequest)
+	if !ok {
+		return utils.WrapErr(errors.New("request not found")), nil
+	}
+
+	awsReq.Body, err = json.Marshal(llamaReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "marshal request")), nil
+	}
+
+	awsResp, err := awsCli.InvokeModelWithResponseStream(c.Request.Context(), awsReq)
+	if err != nil {
+		return utils.WrapErr(errors.Wrap(err, "InvokeModelWithResponseStream")), nil
+	}
+	stream := awsResp.GetStream()
+	defer stream.Close()
+
+	c.Writer.Header().Set("Content-Type", "text/event-stream")
+	var usage relaymodel.Usage
+	c.Stream(func(w io.Writer) bool {
+		event, ok := <-stream.Events()
+		if !ok {
+			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
+			return false
+		}
+
+		switch v := event.(type) {
+		case *types.ResponseStreamMemberChunk:
+			var llamaResp StreamResponse
+			err := json.NewDecoder(bytes.NewReader(v.Value.Bytes)).Decode(&llamaResp)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				return false
+			}
+
+			if llamaResp.PromptTokenCount > 0 {
+				usage.PromptTokens = llamaResp.PromptTokenCount
+			}
+			if llamaResp.StopReason == "stop" {
+				usage.CompletionTokens = llamaResp.GenerationTokenCount
+				usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
+			}
+			response := StreamResponseLlama2OpenAI(&llamaResp)
+			response.Id = fmt.Sprintf("chatcmpl-%s", random.GetUUID())
+			response.Model = c.GetString(ctxkey.OriginalModel)
+			response.Created = createdTime
+			jsonStr, err := json.Marshal(response)
+			if err != nil {
+				logger.SysError("error marshalling stream response: " + err.Error())
+				return true
+			}
+			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
+			return true
+		case *types.UnknownUnionMember:
+			fmt.Println("unknown tag:", v.Tag)
+			return false
+		default:
+			fmt.Println("union is nil or unknown type")
+			return false
+		}
+	})
+
+	return nil, &usage
+}
+
+func StreamResponseLlama2OpenAI(llamaResponse *StreamResponse) *openai.ChatCompletionsStreamResponse {
+	var choice openai.ChatCompletionsStreamResponseChoice
+	choice.Delta.Content = llamaResponse.Generation
+	choice.Delta.Role = "assistant"
+	finishReason := llamaResponse.StopReason
+	if finishReason != "null" {
+		choice.FinishReason = &finishReason
+	}
+	var openaiResponse openai.ChatCompletionsStreamResponse
+	openaiResponse.Object = "chat.completion.chunk"
+	openaiResponse.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
+	return &openaiResponse
+}
diff --git a/relay/adaptor/aws/llama3/main_test.go b/relay/adaptor/aws/llama3/main_test.go
new file mode 100644
index 00000000..d539eee8
--- /dev/null
+++ b/relay/adaptor/aws/llama3/main_test.go
@@ -0,0 +1,45 @@
+package aws_test
+
+import (
+	"testing"
+
+	aws "github.com/songquanpeng/one-api/relay/adaptor/aws/llama3"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestRenderPrompt(t *testing.T) {
+	messages := []relaymodel.Message{
+		{
+			Role:    "user",
+			Content: "What's your name?",
+		},
+	}
+	prompt := aws.RenderPrompt(messages)
+	expected := `<|begin_of_text|><|start_header_id|>user<|end_header_id|>What's your name?<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+`
+	assert.Equal(t, expected, prompt)
+
+	messages = []relaymodel.Message{
+		{
+			Role:    "system",
+			Content: "Your name is Kat. You are a detective.",
+		},
+		{
+			Role:    "user",
+			Content: "What's your name?",
+		},
+		{
+			Role:    "assistant",
+			Content: "Kat",
+		},
+		{
+			Role:    "user",
+			Content: "What's your job?",
+		},
+	}
+	prompt = aws.RenderPrompt(messages)
+	expected = `<|begin_of_text|><|start_header_id|>system<|end_header_id|>Your name is Kat. You are a detective.<|eot_id|><|start_header_id|>user<|end_header_id|>What's your name?<|eot_id|><|start_header_id|>assistant<|end_header_id|>Kat<|eot_id|><|start_header_id|>user<|end_header_id|>What's your job?<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+`
+	assert.Equal(t, expected, prompt)
+}
diff --git a/relay/adaptor/aws/llama3/model.go b/relay/adaptor/aws/llama3/model.go
new file mode 100644
index 00000000..6cb64cde
--- /dev/null
+++ b/relay/adaptor/aws/llama3/model.go
@@ -0,0 +1,29 @@
+package aws
+
+// Request is the request to AWS Llama3
+//
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-meta.html
+type Request struct {
+	Prompt      string   `json:"prompt"`
+	MaxGenLen   int      `json:"max_gen_len,omitempty"`
+	Temperature *float64 `json:"temperature,omitempty"`
+	TopP        *float64 `json:"top_p,omitempty"`
+}
+
+// Response is the response from AWS Llama3
+//
+// https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-meta.html
+type Response struct {
+	Generation           string `json:"generation"`
+	PromptTokenCount     int    `json:"prompt_token_count"`
+	GenerationTokenCount int    `json:"generation_token_count"`
+	StopReason           string `json:"stop_reason"`
+}
+
+// {'generation': 'Hi', 'prompt_token_count': 15, 'generation_token_count': 1, 'stop_reason': None}
+type StreamResponse struct {
+	Generation           string `json:"generation"`
+	PromptTokenCount     int    `json:"prompt_token_count"`
+	GenerationTokenCount int    `json:"generation_token_count"`
+	StopReason           string `json:"stop_reason"`
+}
diff --git a/relay/adaptor/aws/registry.go b/relay/adaptor/aws/registry.go
new file mode 100644
index 00000000..5f655480
--- /dev/null
+++ b/relay/adaptor/aws/registry.go
@@ -0,0 +1,39 @@
+package aws
+
+import (
+	claude "github.com/songquanpeng/one-api/relay/adaptor/aws/claude"
+	llama3 "github.com/songquanpeng/one-api/relay/adaptor/aws/llama3"
+	"github.com/songquanpeng/one-api/relay/adaptor/aws/utils"
+)
+
+type AwsModelType int
+
+const (
+	AwsClaude AwsModelType = iota + 1
+	AwsLlama3
+)
+
+var (
+	adaptors = map[string]AwsModelType{}
+)
+
+func init() {
+	for model := range claude.AwsModelIDMap {
+		adaptors[model] = AwsClaude
+	}
+	for model := range llama3.AwsModelIDMap {
+		adaptors[model] = AwsLlama3
+	}
+}
+
+func GetAdaptor(model string) utils.AwsAdapter {
+	adaptorType := adaptors[model]
+	switch adaptorType {
+	case AwsClaude:
+		return &claude.Adaptor{}
+	case AwsLlama3:
+		return &llama3.Adaptor{}
+	default:
+		return nil
+	}
+}
diff --git a/relay/adaptor/aws/utils/adaptor.go b/relay/adaptor/aws/utils/adaptor.go
new file mode 100644
index 00000000..4cb880f2
--- /dev/null
+++ b/relay/adaptor/aws/utils/adaptor.go
@@ -0,0 +1,51 @@
+package utils
+
+import (
+	"errors"
+	"io"
+	"net/http"
+
+	"github.com/aws/aws-sdk-go-v2/aws"
+	"github.com/aws/aws-sdk-go-v2/credentials"
+	"github.com/aws/aws-sdk-go-v2/service/bedrockruntime"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type AwsAdapter interface {
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	DoResponse(c *gin.Context, awsCli *bedrockruntime.Client, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+}
+
+type Adaptor struct {
+	Meta      *meta.Meta
+	AwsClient *bedrockruntime.Client
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.Meta = meta
+	a.AwsClient = bedrockruntime.New(bedrockruntime.Options{
+		Region:      meta.Config.Region,
+		Credentials: aws.NewCredentialsCache(credentials.NewStaticCredentialsProvider(meta.Config.AK, meta.Config.SK, "")),
+	})
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return "", nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return nil, nil
+}
diff --git a/relay/adaptor/aws/utils/utils.go b/relay/adaptor/aws/utils/utils.go
new file mode 100644
index 00000000..669dc628
--- /dev/null
+++ b/relay/adaptor/aws/utils/utils.go
@@ -0,0 +1,16 @@
+package utils
+
+import (
+	"net/http"
+
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+func WrapErr(err error) *relaymodel.ErrorWithStatusCode {
+	return &relaymodel.ErrorWithStatusCode{
+		StatusCode: http.StatusInternalServerError,
+		Error: relaymodel.Error{
+			Message: err.Error(),
+		},
+	}
+}
diff --git a/relay/channel/baichuan/constants.go b/relay/adaptor/baichuan/constants.go
similarity index 100%
rename from relay/channel/baichuan/constants.go
rename to relay/adaptor/baichuan/constants.go
diff --git a/relay/channel/baidu/adaptor.go b/relay/adaptor/baidu/adaptor.go
similarity index 63%
rename from relay/channel/baidu/adaptor.go
rename to relay/adaptor/baidu/adaptor.go
index 2d2e24f6..15306b95 100644
--- a/relay/channel/baidu/adaptor.go
+++ b/relay/adaptor/baidu/adaptor.go
@@ -3,25 +3,25 @@ package baidu
 import (
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 	"io"
 	"net/http"
 	"strings"
 
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 )
 
 type Adaptor struct {
 }
 
-func (a *Adaptor) Init(meta *util.RelayMeta) {
+func (a *Adaptor) Init(meta *meta.Meta) {
 
 }
 
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/clntwmv7t
 	suffix := "chat/"
 	if strings.HasPrefix(meta.ActualModelName, "Embedding") {
@@ -38,16 +38,34 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 		suffix += "completions_pro"
 	case "ERNIE-Bot-4":
 		suffix += "completions_pro"
-	case "ERNIE-3.5-8K":
-		suffix += "completions"
-	case "ERNIE-Bot-8K":
-		suffix += "ernie_bot_8k"
 	case "ERNIE-Bot":
 		suffix += "completions"
-	case "ERNIE-Speed":
-		suffix += "ernie_speed"
 	case "ERNIE-Bot-turbo":
 		suffix += "eb-instant"
+	case "ERNIE-Speed":
+		suffix += "ernie_speed"
+	case "ERNIE-4.0-8K":
+		suffix += "completions_pro"
+	case "ERNIE-3.5-8K":
+		suffix += "completions"
+	case "ERNIE-3.5-8K-0205":
+		suffix += "ernie-3.5-8k-0205"
+	case "ERNIE-3.5-8K-1222":
+		suffix += "ernie-3.5-8k-1222"
+	case "ERNIE-Bot-8K":
+		suffix += "ernie_bot_8k"
+	case "ERNIE-3.5-4K-0205":
+		suffix += "ernie-3.5-4k-0205"
+	case "ERNIE-Speed-8K":
+		suffix += "ernie_speed"
+	case "ERNIE-Speed-128K":
+		suffix += "ernie-speed-128k"
+	case "ERNIE-Lite-8K-0922":
+		suffix += "eb-instant"
+	case "ERNIE-Lite-8K-0308":
+		suffix += "ernie-lite-8k"
+	case "ERNIE-Tiny-8K":
+		suffix += "ernie-tiny-8k"
 	case "BLOOMZ-7B":
 		suffix += "bloomz_7b1"
 	case "Embedding-V1":
@@ -59,7 +77,7 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 	case "tao-8k":
 		suffix += "tao_8k"
 	default:
-		suffix += meta.ActualModelName
+		suffix += strings.ToLower(meta.ActualModelName)
 	}
 	fullRequestURL := fmt.Sprintf("%s/rpc/2.0/ai_custom/v1/wenxinworkshop/%s", meta.BaseURL, suffix)
 	var accessToken string
@@ -71,8 +89,8 @@ func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
 	return fullRequestURL, nil
 }
 
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
 	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
 	return nil
 }
@@ -82,7 +100,7 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 		return nil, errors.New("request is nil")
 	}
 	switch relayMode {
-	case constant.RelayModeEmbeddings:
+	case relaymode.Embeddings:
 		baiduEmbeddingRequest := ConvertEmbeddingRequest(*request)
 		return baiduEmbeddingRequest, nil
 	default:
@@ -91,16 +109,23 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	}
 }
 
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
 	if meta.IsStream {
 		err, usage = StreamHandler(c, resp)
 	} else {
 		switch meta.Mode {
-		case constant.RelayModeEmbeddings:
+		case relaymode.Embeddings:
 			err, usage = EmbeddingHandler(c, resp)
 		default:
 			err, usage = Handler(c, resp)
diff --git a/relay/adaptor/baidu/constants.go b/relay/adaptor/baidu/constants.go
new file mode 100644
index 00000000..f952adc6
--- /dev/null
+++ b/relay/adaptor/baidu/constants.go
@@ -0,0 +1,20 @@
+package baidu
+
+var ModelList = []string{
+	"ERNIE-4.0-8K",
+	"ERNIE-3.5-8K",
+	"ERNIE-3.5-8K-0205",
+	"ERNIE-3.5-8K-1222",
+	"ERNIE-Bot-8K",
+	"ERNIE-3.5-4K-0205",
+	"ERNIE-Speed-8K",
+	"ERNIE-Speed-128K",
+	"ERNIE-Lite-8K-0922",
+	"ERNIE-Lite-8K-0308",
+	"ERNIE-Tiny-8K",
+	"BLOOMZ-7B",
+	"Embedding-V1",
+	"bge-large-zh",
+	"bge-large-en",
+	"tao-8k",
+}
diff --git a/relay/channel/baidu/main.go b/relay/adaptor/baidu/main.go
similarity index 82%
rename from relay/channel/baidu/main.go
rename to relay/adaptor/baidu/main.go
index 9ca9e47d..ac8a5625 100644
--- a/relay/channel/baidu/main.go
+++ b/relay/adaptor/baidu/main.go
@@ -5,18 +5,20 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
+	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
 	"sync"
 	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/client"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 // https://cloud.baidu.com/doc/WENXINWORKSHOP/s/flfmc9do2
@@ -33,9 +35,9 @@ type Message struct {
 
 type ChatRequest struct {
 	Messages        []Message `json:"messages"`
-	Temperature     float64   `json:"temperature,omitempty"`
-	TopP            float64   `json:"top_p,omitempty"`
-	PenaltyScore    float64   `json:"penalty_score,omitempty"`
+	Temperature     *float64  `json:"temperature,omitempty"`
+	TopP            *float64  `json:"top_p,omitempty"`
+	PenaltyScore    *float64  `json:"penalty_score,omitempty"`
 	Stream          bool      `json:"stream,omitempty"`
 	System          string    `json:"system,omitempty"`
 	DisableSearch   bool      `json:"disable_search,omitempty"`
@@ -137,59 +139,41 @@ func embeddingResponseBaidu2OpenAI(response *EmbeddingResponse) *openai.Embeddin
 func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var usage model.Usage
 	scanner := bufio.NewScanner(resp.Body)
-	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
-		if atEOF && len(data) == 0 {
-			return 0, nil, nil
-		}
-		if i := strings.Index(string(data), "\n"); i >= 0 {
-			return i + 1, data[0:i], nil
-		}
-		if atEOF {
-			return len(data), data, nil
-		}
-		return 0, nil, nil
-	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 6 { // ignore blank line or wrong format
-				continue
-			}
-			data = data[6:]
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+	scanner.Split(bufio.ScanLines)
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			var baiduResponse ChatStreamResponse
-			err := json.Unmarshal([]byte(data), &baiduResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			if baiduResponse.Usage.TotalTokens != 0 {
-				usage.TotalTokens = baiduResponse.Usage.TotalTokens
-				usage.PromptTokens = baiduResponse.Usage.PromptTokens
-				usage.CompletionTokens = baiduResponse.Usage.TotalTokens - baiduResponse.Usage.PromptTokens
-			}
-			response := streamResponseBaidu2OpenAI(&baiduResponse)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 6 {
+			continue
 		}
-	})
+		data = data[6:]
+
+		var baiduResponse ChatStreamResponse
+		err := json.Unmarshal([]byte(data), &baiduResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+		if baiduResponse.Usage.TotalTokens != 0 {
+			usage.TotalTokens = baiduResponse.Usage.TotalTokens
+			usage.PromptTokens = baiduResponse.Usage.PromptTokens
+			usage.CompletionTokens = baiduResponse.Usage.TotalTokens - baiduResponse.Usage.PromptTokens
+		}
+		response := streamResponseBaidu2OpenAI(&baiduResponse)
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
@@ -305,7 +289,7 @@ func getBaiduAccessTokenHelper(apiKey string) (*AccessToken, error) {
 	}
 	req.Header.Add("Content-Type", "application/json")
 	req.Header.Add("Accept", "application/json")
-	res, err := util.ImpatientHTTPClient.Do(req)
+	res, err := client.ImpatientHTTPClient.Do(req)
 	if err != nil {
 		return nil, err
 	}
diff --git a/relay/channel/baidu/model.go b/relay/adaptor/baidu/model.go
similarity index 100%
rename from relay/channel/baidu/model.go
rename to relay/adaptor/baidu/model.go
diff --git a/relay/adaptor/cloudflare/adaptor.go b/relay/adaptor/cloudflare/adaptor.go
new file mode 100644
index 00000000..97e3dbb2
--- /dev/null
+++ b/relay/adaptor/cloudflare/adaptor.go
@@ -0,0 +1,100 @@
+package cloudflare
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+type Adaptor struct {
+	meta *meta.Meta
+}
+
+// ConvertImageRequest implements adaptor.Adaptor.
+func (*Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertImageRequest implements adaptor.Adaptor.
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
+}
+
+// WorkerAI cannot be used across accounts with AIGateWay
+// https://developers.cloudflare.com/ai-gateway/providers/workersai/#openai-compatible-endpoints
+// https://gateway.ai.cloudflare.com/v1/{account_id}/{gateway_id}/workers-ai
+func (a *Adaptor) isAIGateWay(baseURL string) bool {
+	return strings.HasPrefix(baseURL, "https://gateway.ai.cloudflare.com") && strings.HasSuffix(baseURL, "/workers-ai")
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	isAIGateWay := a.isAIGateWay(meta.BaseURL)
+	var urlPrefix string
+	if isAIGateWay {
+		urlPrefix = meta.BaseURL
+	} else {
+		urlPrefix = fmt.Sprintf("%s/client/v4/accounts/%s/ai", meta.BaseURL, meta.Config.UserID)
+	}
+
+	switch meta.Mode {
+	case relaymode.ChatCompletions:
+		return fmt.Sprintf("%s/v1/chat/completions", urlPrefix), nil
+	case relaymode.Embeddings:
+		return fmt.Sprintf("%s/v1/embeddings", urlPrefix), nil
+	default:
+		if isAIGateWay {
+			return fmt.Sprintf("%s/%s", urlPrefix, meta.ActualModelName), nil
+		}
+		return fmt.Sprintf("%s/run/%s", urlPrefix, meta.ActualModelName), nil
+	}
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	switch relayMode {
+	case relaymode.Completions:
+		return ConvertCompletionsRequest(*request), nil
+	case relaymode.ChatCompletions, relaymode.Embeddings:
+		return request, nil
+	default:
+		return nil, errors.New("not implemented")
+	}
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	} else {
+		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "cloudflare"
+}
diff --git a/relay/adaptor/cloudflare/constant.go b/relay/adaptor/cloudflare/constant.go
new file mode 100644
index 00000000..54052aa6
--- /dev/null
+++ b/relay/adaptor/cloudflare/constant.go
@@ -0,0 +1,37 @@
+package cloudflare
+
+var ModelList = []string{
+	"@cf/meta/llama-3.1-8b-instruct",
+	"@cf/meta/llama-2-7b-chat-fp16",
+	"@cf/meta/llama-2-7b-chat-int8",
+	"@cf/mistral/mistral-7b-instruct-v0.1",
+	"@hf/thebloke/deepseek-coder-6.7b-base-awq",
+	"@hf/thebloke/deepseek-coder-6.7b-instruct-awq",
+	"@cf/deepseek-ai/deepseek-math-7b-base",
+	"@cf/deepseek-ai/deepseek-math-7b-instruct",
+	"@cf/thebloke/discolm-german-7b-v1-awq",
+	"@cf/tiiuae/falcon-7b-instruct",
+	"@cf/google/gemma-2b-it-lora",
+	"@hf/google/gemma-7b-it",
+	"@cf/google/gemma-7b-it-lora",
+	"@hf/nousresearch/hermes-2-pro-mistral-7b",
+	"@hf/thebloke/llama-2-13b-chat-awq",
+	"@cf/meta-llama/llama-2-7b-chat-hf-lora",
+	"@cf/meta/llama-3-8b-instruct",
+	"@hf/thebloke/llamaguard-7b-awq",
+	"@hf/thebloke/mistral-7b-instruct-v0.1-awq",
+	"@hf/mistralai/mistral-7b-instruct-v0.2",
+	"@cf/mistral/mistral-7b-instruct-v0.2-lora",
+	"@hf/thebloke/neural-chat-7b-v3-1-awq",
+	"@cf/openchat/openchat-3.5-0106",
+	"@hf/thebloke/openhermes-2.5-mistral-7b-awq",
+	"@cf/microsoft/phi-2",
+	"@cf/qwen/qwen1.5-0.5b-chat",
+	"@cf/qwen/qwen1.5-1.8b-chat",
+	"@cf/qwen/qwen1.5-14b-chat-awq",
+	"@cf/qwen/qwen1.5-7b-chat-awq",
+	"@cf/defog/sqlcoder-7b-2",
+	"@hf/nexusflow/starling-lm-7b-beta",
+	"@cf/tinyllama/tinyllama-1.1b-chat-v1.0",
+	"@hf/thebloke/zephyr-7b-beta-awq",
+}
diff --git a/relay/adaptor/cloudflare/main.go b/relay/adaptor/cloudflare/main.go
new file mode 100644
index 00000000..980a2891
--- /dev/null
+++ b/relay/adaptor/cloudflare/main.go
@@ -0,0 +1,115 @@
+package cloudflare
+
+import (
+	"bufio"
+	"encoding/json"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/render"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+func ConvertCompletionsRequest(textRequest model.GeneralOpenAIRequest) *Request {
+	p, _ := textRequest.Prompt.(string)
+	return &Request{
+		Prompt:      p,
+		MaxTokens:   textRequest.MaxTokens,
+		Stream:      textRequest.Stream,
+		Temperature: textRequest.Temperature,
+	}
+}
+
+func StreamHandler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(bufio.ScanLines)
+
+	common.SetEventStreamHeaders(c)
+	id := helper.GetResponseID(c)
+	responseModel := c.GetString(ctxkey.OriginalModel)
+	var responseText string
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < len("data: ") {
+			continue
+		}
+		data = strings.TrimPrefix(data, "data: ")
+		data = strings.TrimSuffix(data, "\r")
+
+		if data == "[DONE]" {
+			break
+		}
+
+		var response openai.ChatCompletionsStreamResponse
+		err := json.Unmarshal([]byte(data), &response)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+		for _, v := range response.Choices {
+			v.Delta.Role = "assistant"
+			responseText += v.Delta.StringContent()
+		}
+		response.Id = id
+		response.Model = modelName
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
+	err := resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	usage := openai.ResponseText2Usage(responseText, responseModel, promptTokens)
+	return nil, usage
+}
+
+func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	var response openai.TextResponse
+	err = json.Unmarshal(responseBody, &response)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	response.Model = modelName
+	var responseText string
+	for _, v := range response.Choices {
+		responseText += v.Message.Content.(string)
+	}
+	usage := openai.ResponseText2Usage(responseText, modelName, promptTokens)
+	response.Usage = *usage
+	response.Id = helper.GetResponseID(c)
+	jsonResponse, err := json.Marshal(response)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, _ = c.Writer.Write(jsonResponse)
+	return nil, usage
+}
diff --git a/relay/adaptor/cloudflare/model.go b/relay/adaptor/cloudflare/model.go
new file mode 100644
index 00000000..8e382ba7
--- /dev/null
+++ b/relay/adaptor/cloudflare/model.go
@@ -0,0 +1,13 @@
+package cloudflare
+
+import "github.com/songquanpeng/one-api/relay/model"
+
+type Request struct {
+	Messages    []model.Message `json:"messages,omitempty"`
+	Lora        string          `json:"lora,omitempty"`
+	MaxTokens   int             `json:"max_tokens,omitempty"`
+	Prompt      string          `json:"prompt,omitempty"`
+	Raw         bool            `json:"raw,omitempty"`
+	Stream      bool            `json:"stream,omitempty"`
+	Temperature *float64        `json:"temperature,omitempty"`
+}
diff --git a/relay/adaptor/cohere/adaptor.go b/relay/adaptor/cohere/adaptor.go
new file mode 100644
index 00000000..6fdb1b04
--- /dev/null
+++ b/relay/adaptor/cohere/adaptor.go
@@ -0,0 +1,64 @@
+package cohere
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type Adaptor struct{}
+
+// ConvertImageRequest implements adaptor.Adaptor.
+func (*Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertImageRequest implements adaptor.Adaptor.
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return fmt.Sprintf("%s/v1/chat", meta.BaseURL), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return ConvertRequest(*request), nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "Cohere"
+}
diff --git a/relay/adaptor/cohere/constant.go b/relay/adaptor/cohere/constant.go
new file mode 100644
index 00000000..9e70652c
--- /dev/null
+++ b/relay/adaptor/cohere/constant.go
@@ -0,0 +1,14 @@
+package cohere
+
+var ModelList = []string{
+	"command", "command-nightly",
+	"command-light", "command-light-nightly",
+	"command-r", "command-r-plus",
+}
+
+func init() {
+	num := len(ModelList)
+	for i := 0; i < num; i++ {
+		ModelList = append(ModelList, ModelList[i]+"-internet")
+	}
+}
diff --git a/relay/adaptor/cohere/main.go b/relay/adaptor/cohere/main.go
new file mode 100644
index 00000000..736c5a8d
--- /dev/null
+++ b/relay/adaptor/cohere/main.go
@@ -0,0 +1,228 @@
+package cohere
+
+import (
+	"bufio"
+	"encoding/json"
+	"fmt"
+	"github.com/songquanpeng/one-api/common/render"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var (
+	WebSearchConnector = Connector{ID: "web-search"}
+)
+
+func stopReasonCohere2OpenAI(reason *string) string {
+	if reason == nil {
+		return ""
+	}
+	switch *reason {
+	case "COMPLETE":
+		return "stop"
+	default:
+		return *reason
+	}
+}
+
+func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
+	cohereRequest := Request{
+		Model:            textRequest.Model,
+		Message:          "",
+		MaxTokens:        textRequest.MaxTokens,
+		Temperature:      textRequest.Temperature,
+		P:                textRequest.TopP,
+		K:                textRequest.TopK,
+		Stream:           textRequest.Stream,
+		FrequencyPenalty: textRequest.FrequencyPenalty,
+		PresencePenalty:  textRequest.PresencePenalty,
+		Seed:             int(textRequest.Seed),
+	}
+	if cohereRequest.Model == "" {
+		cohereRequest.Model = "command-r"
+	}
+	if strings.HasSuffix(cohereRequest.Model, "-internet") {
+		cohereRequest.Model = strings.TrimSuffix(cohereRequest.Model, "-internet")
+		cohereRequest.Connectors = append(cohereRequest.Connectors, WebSearchConnector)
+	}
+	for _, message := range textRequest.Messages {
+		if message.Role == "user" {
+			cohereRequest.Message = message.Content.(string)
+		} else {
+			var role string
+			if message.Role == "assistant" {
+				role = "CHATBOT"
+			} else if message.Role == "system" {
+				role = "SYSTEM"
+			} else {
+				role = "USER"
+			}
+			cohereRequest.ChatHistory = append(cohereRequest.ChatHistory, ChatMessage{
+				Role:    role,
+				Message: message.Content.(string),
+			})
+		}
+	}
+	return &cohereRequest
+}
+
+func StreamResponseCohere2OpenAI(cohereResponse *StreamResponse) (*openai.ChatCompletionsStreamResponse, *Response) {
+	var response *Response
+	var responseText string
+	var finishReason string
+
+	switch cohereResponse.EventType {
+	case "stream-start":
+		return nil, nil
+	case "text-generation":
+		responseText += cohereResponse.Text
+	case "stream-end":
+		usage := cohereResponse.Response.Meta.Tokens
+		response = &Response{
+			Meta: Meta{
+				Tokens: Usage{
+					InputTokens:  usage.InputTokens,
+					OutputTokens: usage.OutputTokens,
+				},
+			},
+		}
+		finishReason = *cohereResponse.Response.FinishReason
+	default:
+		return nil, nil
+	}
+
+	var choice openai.ChatCompletionsStreamResponseChoice
+	choice.Delta.Content = responseText
+	choice.Delta.Role = "assistant"
+	if finishReason != "" {
+		choice.FinishReason = &finishReason
+	}
+	var openaiResponse openai.ChatCompletionsStreamResponse
+	openaiResponse.Object = "chat.completion.chunk"
+	openaiResponse.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
+	return &openaiResponse, response
+}
+
+func ResponseCohere2OpenAI(cohereResponse *Response) *openai.TextResponse {
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: model.Message{
+			Role:    "assistant",
+			Content: cohereResponse.Text,
+			Name:    nil,
+		},
+		FinishReason: stopReasonCohere2OpenAI(cohereResponse.FinishReason),
+	}
+	fullTextResponse := openai.TextResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", cohereResponse.ResponseID),
+		Model:   "model",
+		Object:  "chat.completion",
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+	}
+	return &fullTextResponse
+}
+
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	createdTime := helper.GetTimestamp()
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(bufio.ScanLines)
+
+	common.SetEventStreamHeaders(c)
+	var usage model.Usage
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		data = strings.TrimSuffix(data, "\r")
+
+		var cohereResponse StreamResponse
+		err := json.Unmarshal([]byte(data), &cohereResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		response, meta := StreamResponseCohere2OpenAI(&cohereResponse)
+		if meta != nil {
+			usage.PromptTokens += meta.Meta.Tokens.InputTokens
+			usage.CompletionTokens += meta.Meta.Tokens.OutputTokens
+			continue
+		}
+		if response == nil {
+			continue
+		}
+
+		response.Id = fmt.Sprintf("chatcmpl-%d", createdTime)
+		response.Model = c.GetString("original_model")
+		response.Created = createdTime
+
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
+	err := resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	return nil, &usage
+}
+
+func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *model.Usage) {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	var cohereResponse Response
+	err = json.Unmarshal(responseBody, &cohereResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	if cohereResponse.ResponseID == "" {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: cohereResponse.Message,
+				Type:    cohereResponse.Message,
+				Param:   "",
+				Code:    resp.StatusCode,
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+	fullTextResponse := ResponseCohere2OpenAI(&cohereResponse)
+	fullTextResponse.Model = modelName
+	usage := model.Usage{
+		PromptTokens:     cohereResponse.Meta.Tokens.InputTokens,
+		CompletionTokens: cohereResponse.Meta.Tokens.OutputTokens,
+		TotalTokens:      cohereResponse.Meta.Tokens.InputTokens + cohereResponse.Meta.Tokens.OutputTokens,
+	}
+	fullTextResponse.Usage = usage
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &usage
+}
diff --git a/relay/adaptor/cohere/model.go b/relay/adaptor/cohere/model.go
new file mode 100644
index 00000000..3a8bc99d
--- /dev/null
+++ b/relay/adaptor/cohere/model.go
@@ -0,0 +1,147 @@
+package cohere
+
+type Request struct {
+	Message          string        `json:"message" required:"true"`
+	Model            string        `json:"model,omitempty"`  // 默认值为"command-r"
+	Stream           bool          `json:"stream,omitempty"` // 默认值为false
+	Preamble         string        `json:"preamble,omitempty"`
+	ChatHistory      []ChatMessage `json:"chat_history,omitempty"`
+	ConversationID   string        `json:"conversation_id,omitempty"`
+	PromptTruncation string        `json:"prompt_truncation,omitempty"` // 默认值为"AUTO"
+	Connectors       []Connector   `json:"connectors,omitempty"`
+	Documents        []Document    `json:"documents,omitempty"`
+	Temperature      *float64      `json:"temperature,omitempty"` // 默认值为0.3
+	MaxTokens        int           `json:"max_tokens,omitempty"`
+	MaxInputTokens   int           `json:"max_input_tokens,omitempty"`
+	K                int           `json:"k,omitempty"` // 默认值为0
+	P                *float64      `json:"p,omitempty"` // 默认值为0.75
+	Seed             int           `json:"seed,omitempty"`
+	StopSequences    []string      `json:"stop_sequences,omitempty"`
+	FrequencyPenalty *float64      `json:"frequency_penalty,omitempty"` // 默认值为0.0
+	PresencePenalty  *float64      `json:"presence_penalty,omitempty"`  // 默认值为0.0
+	Tools            []Tool        `json:"tools,omitempty"`
+	ToolResults      []ToolResult  `json:"tool_results,omitempty"`
+}
+
+type ChatMessage struct {
+	Role    string `json:"role" required:"true"`
+	Message string `json:"message" required:"true"`
+}
+
+type Tool struct {
+	Name                 string                   `json:"name" required:"true"`
+	Description          string                   `json:"description" required:"true"`
+	ParameterDefinitions map[string]ParameterSpec `json:"parameter_definitions"`
+}
+
+type ParameterSpec struct {
+	Description string `json:"description"`
+	Type        string `json:"type" required:"true"`
+	Required    bool   `json:"required"`
+}
+
+type ToolResult struct {
+	Call    ToolCall                 `json:"call"`
+	Outputs []map[string]interface{} `json:"outputs"`
+}
+
+type ToolCall struct {
+	Name       string                 `json:"name" required:"true"`
+	Parameters map[string]interface{} `json:"parameters" required:"true"`
+}
+
+type StreamResponse struct {
+	IsFinished    bool            `json:"is_finished"`
+	EventType     string          `json:"event_type"`
+	GenerationID  string          `json:"generation_id,omitempty"`
+	SearchQueries []*SearchQuery  `json:"search_queries,omitempty"`
+	SearchResults []*SearchResult `json:"search_results,omitempty"`
+	Documents     []*Document     `json:"documents,omitempty"`
+	Text          string          `json:"text,omitempty"`
+	Citations     []*Citation     `json:"citations,omitempty"`
+	Response      *Response       `json:"response,omitempty"`
+	FinishReason  string          `json:"finish_reason,omitempty"`
+}
+
+type SearchQuery struct {
+	Text         string `json:"text"`
+	GenerationID string `json:"generation_id"`
+}
+
+type SearchResult struct {
+	SearchQuery *SearchQuery `json:"search_query"`
+	DocumentIDs []string     `json:"document_ids"`
+	Connector   *Connector   `json:"connector"`
+}
+
+type Connector struct {
+	ID string `json:"id"`
+}
+
+type Document struct {
+	ID        string `json:"id"`
+	Snippet   string `json:"snippet"`
+	Timestamp string `json:"timestamp"`
+	Title     string `json:"title"`
+	URL       string `json:"url"`
+}
+
+type Citation struct {
+	Start       int      `json:"start"`
+	End         int      `json:"end"`
+	Text        string   `json:"text"`
+	DocumentIDs []string `json:"document_ids"`
+}
+
+type Response struct {
+	ResponseID    string          `json:"response_id"`
+	Text          string          `json:"text"`
+	GenerationID  string          `json:"generation_id"`
+	ChatHistory   []*Message      `json:"chat_history"`
+	FinishReason  *string         `json:"finish_reason"`
+	Meta          Meta            `json:"meta"`
+	Citations     []*Citation     `json:"citations"`
+	Documents     []*Document     `json:"documents"`
+	SearchResults []*SearchResult `json:"search_results"`
+	SearchQueries []*SearchQuery  `json:"search_queries"`
+	Message       string          `json:"message"`
+}
+
+type Message struct {
+	Role    string `json:"role"`
+	Message string `json:"message"`
+}
+
+type Version struct {
+	Version string `json:"version"`
+}
+
+type Units struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
+}
+
+type ChatEntry struct {
+	Role    string `json:"role"`
+	Message string `json:"message"`
+}
+
+type Meta struct {
+	APIVersion  APIVersion  `json:"api_version"`
+	BilledUnits BilledUnits `json:"billed_units"`
+	Tokens      Usage       `json:"tokens"`
+}
+
+type APIVersion struct {
+	Version string `json:"version"`
+}
+
+type BilledUnits struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
+}
+
+type Usage struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
+}
diff --git a/relay/channel/common.go b/relay/adaptor/common.go
similarity index 80%
rename from relay/channel/common.go
rename to relay/adaptor/common.go
index c6e1abf2..8953d7a3 100644
--- a/relay/channel/common.go
+++ b/relay/adaptor/common.go
@@ -1,15 +1,16 @@
-package channel
+package adaptor
 
 import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/util"
+	"github.com/songquanpeng/one-api/common/client"
+	"github.com/songquanpeng/one-api/relay/meta"
 	"io"
 	"net/http"
 )
 
-func SetupCommonRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) {
+func SetupCommonRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) {
 	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
 	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
 	if meta.IsStream && c.Request.Header.Get("Accept") == "" {
@@ -17,7 +18,7 @@ func SetupCommonRequestHeader(c *gin.Context, req *http.Request, meta *util.Rela
 	}
 }
 
-func DoRequestHelper(a Adaptor, c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+func DoRequestHelper(a Adaptor, c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
 	fullRequestURL, err := a.GetRequestURL(meta)
 	if err != nil {
 		return nil, fmt.Errorf("get request url failed: %w", err)
@@ -38,7 +39,7 @@ func DoRequestHelper(a Adaptor, c *gin.Context, meta *util.RelayMeta, requestBod
 }
 
 func DoRequest(c *gin.Context, req *http.Request) (*http.Response, error) {
-	resp, err := util.HTTPClient.Do(req)
+	resp, err := client.HTTPClient.Do(req)
 	if err != nil {
 		return nil, err
 	}
diff --git a/relay/adaptor/coze/adaptor.go b/relay/adaptor/coze/adaptor.go
new file mode 100644
index 00000000..44f560e8
--- /dev/null
+++ b/relay/adaptor/coze/adaptor.go
@@ -0,0 +1,75 @@
+package coze
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+)
+
+type Adaptor struct {
+	meta *meta.Meta
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return fmt.Sprintf("%s/open_api/v2/chat", meta.BaseURL), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	request.User = a.meta.Config.UserID
+	return ConvertRequest(*request), nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	var responseText *string
+	if meta.IsStream {
+		err, responseText = StreamHandler(c, resp)
+	} else {
+		err, responseText = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	if responseText != nil {
+		usage = openai.ResponseText2Usage(*responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		usage = &model.Usage{}
+	}
+	usage.PromptTokens = meta.PromptTokens
+	usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "coze"
+}
diff --git a/relay/adaptor/coze/constant/contenttype/define.go b/relay/adaptor/coze/constant/contenttype/define.go
new file mode 100644
index 00000000..69c876bc
--- /dev/null
+++ b/relay/adaptor/coze/constant/contenttype/define.go
@@ -0,0 +1,5 @@
+package contenttype
+
+const (
+	Text = "text"
+)
diff --git a/relay/adaptor/coze/constant/event/define.go b/relay/adaptor/coze/constant/event/define.go
new file mode 100644
index 00000000..c03e8c17
--- /dev/null
+++ b/relay/adaptor/coze/constant/event/define.go
@@ -0,0 +1,7 @@
+package event
+
+const (
+	Message = "message"
+	Done    = "done"
+	Error   = "error"
+)
diff --git a/relay/adaptor/coze/constant/messagetype/define.go b/relay/adaptor/coze/constant/messagetype/define.go
new file mode 100644
index 00000000..6c1c25db
--- /dev/null
+++ b/relay/adaptor/coze/constant/messagetype/define.go
@@ -0,0 +1,6 @@
+package messagetype
+
+const (
+	Answer   = "answer"
+	FollowUp = "follow_up"
+)
diff --git a/relay/adaptor/coze/constants.go b/relay/adaptor/coze/constants.go
new file mode 100644
index 00000000..d20fd875
--- /dev/null
+++ b/relay/adaptor/coze/constants.go
@@ -0,0 +1,3 @@
+package coze
+
+var ModelList = []string{}
diff --git a/relay/adaptor/coze/helper.go b/relay/adaptor/coze/helper.go
new file mode 100644
index 00000000..0396afcb
--- /dev/null
+++ b/relay/adaptor/coze/helper.go
@@ -0,0 +1,10 @@
+package coze
+
+import "github.com/songquanpeng/one-api/relay/adaptor/coze/constant/event"
+
+func event2StopReason(e *string) string {
+	if e == nil || *e == event.Message {
+		return ""
+	}
+	return "stop"
+}
diff --git a/relay/adaptor/coze/main.go b/relay/adaptor/coze/main.go
new file mode 100644
index 00000000..d0402a76
--- /dev/null
+++ b/relay/adaptor/coze/main.go
@@ -0,0 +1,202 @@
+package coze
+
+import (
+	"bufio"
+	"encoding/json"
+	"fmt"
+	"github.com/songquanpeng/one-api/common/render"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/conv"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/coze/constant/messagetype"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+// https://www.coze.com/open
+
+func stopReasonCoze2OpenAI(reason *string) string {
+	if reason == nil {
+		return ""
+	}
+	switch *reason {
+	case "end_turn":
+		return "stop"
+	case "stop_sequence":
+		return "stop"
+	case "max_tokens":
+		return "length"
+	default:
+		return *reason
+	}
+}
+
+func ConvertRequest(textRequest model.GeneralOpenAIRequest) *Request {
+	cozeRequest := Request{
+		Stream: textRequest.Stream,
+		User:   textRequest.User,
+		BotId:  strings.TrimPrefix(textRequest.Model, "bot-"),
+	}
+	for i, message := range textRequest.Messages {
+		if i == len(textRequest.Messages)-1 {
+			cozeRequest.Query = message.StringContent()
+			continue
+		}
+		cozeMessage := Message{
+			Role:    message.Role,
+			Content: message.StringContent(),
+		}
+		cozeRequest.ChatHistory = append(cozeRequest.ChatHistory, cozeMessage)
+	}
+	return &cozeRequest
+}
+
+func StreamResponseCoze2OpenAI(cozeResponse *StreamResponse) (*openai.ChatCompletionsStreamResponse, *Response) {
+	var response *Response
+	var stopReason string
+	var choice openai.ChatCompletionsStreamResponseChoice
+
+	if cozeResponse.Message != nil {
+		if cozeResponse.Message.Type != messagetype.Answer {
+			return nil, nil
+		}
+		choice.Delta.Content = cozeResponse.Message.Content
+	}
+	choice.Delta.Role = "assistant"
+	finishReason := stopReasonCoze2OpenAI(&stopReason)
+	if finishReason != "null" {
+		choice.FinishReason = &finishReason
+	}
+	var openaiResponse openai.ChatCompletionsStreamResponse
+	openaiResponse.Object = "chat.completion.chunk"
+	openaiResponse.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
+	openaiResponse.Id = cozeResponse.ConversationId
+	return &openaiResponse, response
+}
+
+func ResponseCoze2OpenAI(cozeResponse *Response) *openai.TextResponse {
+	var responseText string
+	for _, message := range cozeResponse.Messages {
+		if message.Type == messagetype.Answer {
+			responseText = message.Content
+			break
+		}
+	}
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: model.Message{
+			Role:    "assistant",
+			Content: responseText,
+			Name:    nil,
+		},
+		FinishReason: "stop",
+	}
+	fullTextResponse := openai.TextResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", cozeResponse.ConversationId),
+		Model:   "coze-bot",
+		Object:  "chat.completion",
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+	}
+	return &fullTextResponse
+}
+
+func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *string) {
+	var responseText string
+	createdTime := helper.GetTimestamp()
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(bufio.ScanLines)
+
+	common.SetEventStreamHeaders(c)
+	var modelName string
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 5 || !strings.HasPrefix(data, "data:") {
+			continue
+		}
+		data = strings.TrimPrefix(data, "data:")
+		data = strings.TrimSuffix(data, "\r")
+
+		var cozeResponse StreamResponse
+		err := json.Unmarshal([]byte(data), &cozeResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		response, _ := StreamResponseCoze2OpenAI(&cozeResponse)
+		if response == nil {
+			continue
+		}
+
+		for _, choice := range response.Choices {
+			responseText += conv.AsString(choice.Delta.Content)
+		}
+		response.Model = modelName
+		response.Created = createdTime
+
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
+	err := resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	return nil, &responseText
+}
+
+func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName string) (*model.ErrorWithStatusCode, *string) {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	var cozeResponse Response
+	err = json.Unmarshal(responseBody, &cozeResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	if cozeResponse.Code != 0 {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: cozeResponse.Msg,
+				Code:    cozeResponse.Code,
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+	fullTextResponse := ResponseCoze2OpenAI(&cozeResponse)
+	fullTextResponse.Model = modelName
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	var responseText string
+	if len(fullTextResponse.Choices) > 0 {
+		responseText = fullTextResponse.Choices[0].Message.StringContent()
+	}
+	return nil, &responseText
+}
diff --git a/relay/adaptor/coze/model.go b/relay/adaptor/coze/model.go
new file mode 100644
index 00000000..d0afecfe
--- /dev/null
+++ b/relay/adaptor/coze/model.go
@@ -0,0 +1,38 @@
+package coze
+
+type Message struct {
+	Role        string `json:"role"`
+	Type        string `json:"type"`
+	Content     string `json:"content"`
+	ContentType string `json:"content_type"`
+}
+
+type ErrorInformation struct {
+	Code int    `json:"code"`
+	Msg  string `json:"msg"`
+}
+
+type Request struct {
+	ConversationId string    `json:"conversation_id,omitempty"`
+	BotId          string    `json:"bot_id"`
+	User           string    `json:"user"`
+	Query          string    `json:"query"`
+	ChatHistory    []Message `json:"chat_history,omitempty"`
+	Stream         bool      `json:"stream"`
+}
+
+type Response struct {
+	ConversationId string    `json:"conversation_id,omitempty"`
+	Messages       []Message `json:"messages,omitempty"`
+	Code           int       `json:"code,omitempty"`
+	Msg            string    `json:"msg,omitempty"`
+}
+
+type StreamResponse struct {
+	Event            string            `json:"event,omitempty"`
+	Message          *Message          `json:"message,omitempty"`
+	IsFinish         bool              `json:"is_finish,omitempty"`
+	Index            int               `json:"index,omitempty"`
+	ConversationId   string            `json:"conversation_id,omitempty"`
+	ErrorInformation *ErrorInformation `json:"error_information,omitempty"`
+}
diff --git a/relay/adaptor/deepl/adaptor.go b/relay/adaptor/deepl/adaptor.go
new file mode 100644
index 00000000..d018a096
--- /dev/null
+++ b/relay/adaptor/deepl/adaptor.go
@@ -0,0 +1,73 @@
+package deepl
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+)
+
+type Adaptor struct {
+	meta       *meta.Meta
+	promptText string
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return fmt.Sprintf("%s/v2/translate", meta.BaseURL), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "DeepL-Auth-Key "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	convertedRequest, text := ConvertRequest(*request)
+	a.promptText = text
+	return convertedRequest, nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err = StreamHandler(c, resp, meta.ActualModelName)
+	} else {
+		err = Handler(c, resp, meta.ActualModelName)
+	}
+	promptTokens := len(a.promptText)
+	usage = &model.Usage{
+		PromptTokens: promptTokens,
+		TotalTokens:  promptTokens,
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "deepl"
+}
diff --git a/relay/adaptor/deepl/constants.go b/relay/adaptor/deepl/constants.go
new file mode 100644
index 00000000..6a4f2545
--- /dev/null
+++ b/relay/adaptor/deepl/constants.go
@@ -0,0 +1,9 @@
+package deepl
+
+// https://developers.deepl.com/docs/api-reference/glossaries
+
+var ModelList = []string{
+	"deepl-zh",
+	"deepl-en",
+	"deepl-ja",
+}
diff --git a/relay/adaptor/deepl/helper.go b/relay/adaptor/deepl/helper.go
new file mode 100644
index 00000000..6d3a914b
--- /dev/null
+++ b/relay/adaptor/deepl/helper.go
@@ -0,0 +1,11 @@
+package deepl
+
+import "strings"
+
+func parseLangFromModelName(modelName string) string {
+	parts := strings.Split(modelName, "-")
+	if len(parts) == 1 {
+		return "ZH"
+	}
+	return parts[1]
+}
diff --git a/relay/adaptor/deepl/main.go b/relay/adaptor/deepl/main.go
new file mode 100644
index 00000000..f8bbae14
--- /dev/null
+++ b/relay/adaptor/deepl/main.go
@@ -0,0 +1,137 @@
+package deepl
+
+import (
+	"encoding/json"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/constant/finishreason"
+	"github.com/songquanpeng/one-api/relay/constant/role"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+)
+
+// https://developers.deepl.com/docs/getting-started/your-first-api-request
+
+func ConvertRequest(textRequest model.GeneralOpenAIRequest) (*Request, string) {
+	var text string
+	if len(textRequest.Messages) != 0 {
+		text = textRequest.Messages[len(textRequest.Messages)-1].StringContent()
+	}
+	deeplRequest := Request{
+		TargetLang: parseLangFromModelName(textRequest.Model),
+		Text:       []string{text},
+	}
+	return &deeplRequest, text
+}
+
+func StreamResponseDeepL2OpenAI(deeplResponse *Response) *openai.ChatCompletionsStreamResponse {
+	var choice openai.ChatCompletionsStreamResponseChoice
+	if len(deeplResponse.Translations) != 0 {
+		choice.Delta.Content = deeplResponse.Translations[0].Text
+	}
+	choice.Delta.Role = role.Assistant
+	choice.FinishReason = &constant.StopFinishReason
+	openaiResponse := openai.ChatCompletionsStreamResponse{
+		Object:  constant.StreamObject,
+		Created: helper.GetTimestamp(),
+		Choices: []openai.ChatCompletionsStreamResponseChoice{choice},
+	}
+	return &openaiResponse
+}
+
+func ResponseDeepL2OpenAI(deeplResponse *Response) *openai.TextResponse {
+	var responseText string
+	if len(deeplResponse.Translations) != 0 {
+		responseText = deeplResponse.Translations[0].Text
+	}
+	choice := openai.TextResponseChoice{
+		Index: 0,
+		Message: model.Message{
+			Role:    role.Assistant,
+			Content: responseText,
+			Name:    nil,
+		},
+		FinishReason: finishreason.Stop,
+	}
+	fullTextResponse := openai.TextResponse{
+		Object:  constant.NonStreamObject,
+		Created: helper.GetTimestamp(),
+		Choices: []openai.TextResponseChoice{choice},
+	}
+	return &fullTextResponse
+}
+
+func StreamHandler(c *gin.Context, resp *http.Response, modelName string) *model.ErrorWithStatusCode {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
+	}
+	var deeplResponse Response
+	err = json.Unmarshal(responseBody, &deeplResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError)
+	}
+	fullTextResponse := StreamResponseDeepL2OpenAI(&deeplResponse)
+	fullTextResponse.Model = modelName
+	fullTextResponse.Id = helper.GetResponseID(c)
+	jsonData, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError)
+	}
+	common.SetEventStreamHeaders(c)
+	c.Stream(func(w io.Writer) bool {
+		if jsonData != nil {
+			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonData)})
+			jsonData = nil
+			return true
+		}
+		c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
+		return false
+	})
+	_ = resp.Body.Close()
+	return nil
+}
+
+func Handler(c *gin.Context, resp *http.Response, modelName string) *model.ErrorWithStatusCode {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
+	}
+	var deeplResponse Response
+	err = json.Unmarshal(responseBody, &deeplResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError)
+	}
+	if deeplResponse.Message != "" {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: deeplResponse.Message,
+				Code:    "deepl_error",
+			},
+			StatusCode: resp.StatusCode,
+		}
+	}
+	fullTextResponse := ResponseDeepL2OpenAI(&deeplResponse)
+	fullTextResponse.Model = modelName
+	fullTextResponse.Id = helper.GetResponseID(c)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError)
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil
+}
diff --git a/relay/adaptor/deepl/model.go b/relay/adaptor/deepl/model.go
new file mode 100644
index 00000000..3f823d21
--- /dev/null
+++ b/relay/adaptor/deepl/model.go
@@ -0,0 +1,16 @@
+package deepl
+
+type Request struct {
+	Text       []string `json:"text"`
+	TargetLang string   `json:"target_lang"`
+}
+
+type Translation struct {
+	DetectedSourceLanguage string `json:"detected_source_language,omitempty"`
+	Text                   string `json:"text,omitempty"`
+}
+
+type Response struct {
+	Translations []Translation `json:"translations,omitempty"`
+	Message      string        `json:"message,omitempty"`
+}
diff --git a/relay/adaptor/deepseek/constants.go b/relay/adaptor/deepseek/constants.go
new file mode 100644
index 00000000..ad840bc2
--- /dev/null
+++ b/relay/adaptor/deepseek/constants.go
@@ -0,0 +1,6 @@
+package deepseek
+
+var ModelList = []string{
+	"deepseek-chat",
+	"deepseek-coder",
+}
diff --git a/relay/adaptor/doubao/constants.go b/relay/adaptor/doubao/constants.go
new file mode 100644
index 00000000..dbe819dd
--- /dev/null
+++ b/relay/adaptor/doubao/constants.go
@@ -0,0 +1,13 @@
+package doubao
+
+// https://console.volcengine.com/ark/region:ark+cn-beijing/model
+
+var ModelList = []string{
+	"Doubao-pro-128k",
+	"Doubao-pro-32k",
+	"Doubao-pro-4k",
+	"Doubao-lite-128k",
+	"Doubao-lite-32k",
+	"Doubao-lite-4k",
+	"Doubao-embedding",
+}
diff --git a/relay/adaptor/doubao/main.go b/relay/adaptor/doubao/main.go
new file mode 100644
index 00000000..dd43d06c
--- /dev/null
+++ b/relay/adaptor/doubao/main.go
@@ -0,0 +1,18 @@
+package doubao
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+func GetRequestURL(meta *meta.Meta) (string, error) {
+	switch meta.Mode {
+	case relaymode.ChatCompletions:
+		return fmt.Sprintf("%s/api/v3/chat/completions", meta.BaseURL), nil
+	case relaymode.Embeddings:
+		return fmt.Sprintf("%s/api/v3/embeddings", meta.BaseURL), nil
+	default:
+	}
+	return "", fmt.Errorf("unsupported relay mode %d for doubao", meta.Mode)
+}
diff --git a/relay/adaptor/gemini/adaptor.go b/relay/adaptor/gemini/adaptor.go
new file mode 100644
index 00000000..a86fde40
--- /dev/null
+++ b/relay/adaptor/gemini/adaptor.go
@@ -0,0 +1,101 @@
+package gemini
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/helper"
+	channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	defaultVersion := config.GeminiVersion
+	if meta.ActualModelName == "gemini-2.0-flash-exp" {
+		defaultVersion = "v1beta"
+	}
+
+	version := helper.AssignOrDefault(meta.Config.APIVersion, defaultVersion)
+	action := ""
+	switch meta.Mode {
+	case relaymode.Embeddings:
+		action = "batchEmbedContents"
+	default:
+		action = "generateContent"
+	}
+
+	if meta.IsStream {
+		action = "streamGenerateContent?alt=sse"
+	}
+
+	return fmt.Sprintf("%s/%s/models/%s:%s", meta.BaseURL, version, meta.ActualModelName, action), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	channelhelper.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("x-goog-api-key", meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	switch relayMode {
+	case relaymode.Embeddings:
+		geminiEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return geminiEmbeddingRequest, nil
+	default:
+		geminiRequest := ConvertRequest(*request)
+		return geminiRequest, nil
+	}
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText = StreamHandler(c, resp)
+		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		switch meta.Mode {
+		case relaymode.Embeddings:
+			err, usage = EmbeddingHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "google gemini"
+}
diff --git a/relay/adaptor/gemini/constants.go b/relay/adaptor/gemini/constants.go
new file mode 100644
index 00000000..9d1cbc4a
--- /dev/null
+++ b/relay/adaptor/gemini/constants.go
@@ -0,0 +1,11 @@
+package gemini
+
+// https://ai.google.dev/models/gemini
+
+var ModelList = []string{
+	"gemini-pro", "gemini-1.0-pro",
+	"gemini-1.5-flash", "gemini-1.5-pro",
+	"text-embedding-004", "aqa",
+	"gemini-2.0-flash-exp",
+	"gemini-2.0-flash-thinking-exp",
+}
diff --git a/relay/channel/gemini/main.go b/relay/adaptor/gemini/main.go
similarity index 55%
rename from relay/channel/gemini/main.go
rename to relay/adaptor/gemini/main.go
index c24694c8..3aa4974b 100644
--- a/relay/channel/gemini/main.go
+++ b/relay/adaptor/gemini/main.go
@@ -4,17 +4,21 @@ import (
 	"bufio"
 	"encoding/json"
 	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/songquanpeng/one-api/common/render"
+
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/image"
 	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
-	"strings"
 
 	"github.com/gin-gonic/gin"
 )
@@ -25,6 +29,11 @@ const (
 	VisionMaxImageNum = 16
 )
 
+var mimeTypeMap = map[string]string{
+	"json_object": "application/json",
+	"text":        "text/plain",
+}
+
 // Setting safety to the lowest possible values since Gemini is already powerless enough
 func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 	geminiRequest := ChatRequest{
@@ -46,6 +55,10 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 				Category:  "HARM_CATEGORY_DANGEROUS_CONTENT",
 				Threshold: config.GeminiSafetySetting,
 			},
+			{
+				Category:  "HARM_CATEGORY_CIVIC_INTEGRITY",
+				Threshold: config.GeminiSafetySetting,
+			},
 		},
 		GenerationConfig: ChatGenerationConfig{
 			Temperature:     textRequest.Temperature,
@@ -53,7 +66,26 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 			MaxOutputTokens: textRequest.MaxTokens,
 		},
 	}
-	if textRequest.Functions != nil {
+	if textRequest.ResponseFormat != nil {
+		if mimeType, ok := mimeTypeMap[textRequest.ResponseFormat.Type]; ok {
+			geminiRequest.GenerationConfig.ResponseMimeType = mimeType
+		}
+		if textRequest.ResponseFormat.JsonSchema != nil {
+			geminiRequest.GenerationConfig.ResponseSchema = textRequest.ResponseFormat.JsonSchema.Schema
+			geminiRequest.GenerationConfig.ResponseMimeType = mimeTypeMap["json_object"]
+		}
+	}
+	if textRequest.Tools != nil {
+		functions := make([]model.Function, 0, len(textRequest.Tools))
+		for _, tool := range textRequest.Tools {
+			functions = append(functions, tool.Function)
+		}
+		geminiRequest.Tools = []ChatTools{
+			{
+				FunctionDeclarations: functions,
+			},
+		}
+	} else if textRequest.Functions != nil {
 		geminiRequest.Tools = []ChatTools{
 			{
 				FunctionDeclarations: textRequest.Functions,
@@ -122,6 +154,29 @@ func ConvertRequest(textRequest model.GeneralOpenAIRequest) *ChatRequest {
 	return &geminiRequest
 }
 
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *BatchEmbeddingRequest {
+	inputs := request.ParseInput()
+	requests := make([]EmbeddingRequest, len(inputs))
+	model := fmt.Sprintf("models/%s", request.Model)
+
+	for i, input := range inputs {
+		requests[i] = EmbeddingRequest{
+			Model: model,
+			Content: ChatContent{
+				Parts: []Part{
+					{
+						Text: input,
+					},
+				},
+			},
+		}
+	}
+
+	return &BatchEmbeddingRequest{
+		Requests: requests,
+	}
+}
+
 type ChatResponse struct {
 	Candidates     []ChatCandidate    `json:"candidates"`
 	PromptFeedback ChatPromptFeedback `json:"promptFeedback"`
@@ -153,9 +208,33 @@ type ChatPromptFeedback struct {
 	SafetyRatings []ChatSafetyRating `json:"safetyRatings"`
 }
 
+func getToolCalls(candidate *ChatCandidate) []model.Tool {
+	var toolCalls []model.Tool
+
+	item := candidate.Content.Parts[0]
+	if item.FunctionCall == nil {
+		return toolCalls
+	}
+	argsBytes, err := json.Marshal(item.FunctionCall.Arguments)
+	if err != nil {
+		logger.FatalLog("getToolCalls failed: " + err.Error())
+		return toolCalls
+	}
+	toolCall := model.Tool{
+		Id:   fmt.Sprintf("call_%s", random.GetUUID()),
+		Type: "function",
+		Function: model.Function{
+			Arguments: string(argsBytes),
+			Name:      item.FunctionCall.FunctionName,
+		},
+	}
+	toolCalls = append(toolCalls, toolCall)
+	return toolCalls
+}
+
 func responseGeminiChat2OpenAI(response *ChatResponse) *openai.TextResponse {
 	fullTextResponse := openai.TextResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: make([]openai.TextResponseChoice, 0, len(response.Candidates)),
@@ -164,13 +243,26 @@ func responseGeminiChat2OpenAI(response *ChatResponse) *openai.TextResponse {
 		choice := openai.TextResponseChoice{
 			Index: i,
 			Message: model.Message{
-				Role:    "assistant",
-				Content: "",
+				Role: "assistant",
 			},
 			FinishReason: constant.StopFinishReason,
 		}
 		if len(candidate.Content.Parts) > 0 {
-			choice.Message.Content = candidate.Content.Parts[0].Text
+			if candidate.Content.Parts[0].FunctionCall != nil {
+				choice.Message.ToolCalls = getToolCalls(&candidate)
+			} else {
+				var builder strings.Builder
+				for _, part := range candidate.Content.Parts {
+					if i > 0 {
+						builder.WriteString("\n")
+					}
+					builder.WriteString(part.Text)
+				}
+				choice.Message.Content = builder.String()
+			}
+		} else {
+			choice.Message.Content = ""
+			choice.FinishReason = candidate.FinishReason
 		}
 		fullTextResponse.Choices = append(fullTextResponse.Choices, choice)
 	}
@@ -180,81 +272,80 @@ func responseGeminiChat2OpenAI(response *ChatResponse) *openai.TextResponse {
 func streamResponseGeminiChat2OpenAI(geminiResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = geminiResponse.GetResponseText()
-	choice.FinishReason = &constant.StopFinishReason
+	//choice.FinishReason = &constant.StopFinishReason
 	var response openai.ChatCompletionsStreamResponse
+	response.Id = fmt.Sprintf("chatcmpl-%s", random.GetUUID())
+	response.Created = helper.GetTimestamp()
 	response.Object = "chat.completion.chunk"
 	response.Model = "gemini"
 	response.Choices = []openai.ChatCompletionsStreamResponseChoice{choice}
 	return &response
 }
 
+func embeddingResponseGemini2OpenAI(response *EmbeddingResponse) *openai.EmbeddingResponse {
+	openAIEmbeddingResponse := openai.EmbeddingResponse{
+		Object: "list",
+		Data:   make([]openai.EmbeddingResponseItem, 0, len(response.Embeddings)),
+		Model:  "gemini-embedding",
+		Usage:  model.Usage{TotalTokens: 0},
+	}
+	for _, item := range response.Embeddings {
+		openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
+			Object:    `embedding`,
+			Index:     0,
+			Embedding: item.Values,
+		})
+	}
+	return &openAIEmbeddingResponse
+}
+
 func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, string) {
 	responseText := ""
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
 	scanner := bufio.NewScanner(resp.Body)
-	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
-		if atEOF && len(data) == 0 {
-			return 0, nil, nil
-		}
-		if i := strings.Index(string(data), "\n"); i >= 0 {
-			return i + 1, data[0:i], nil
-		}
-		if atEOF {
-			return len(data), data, nil
-		}
-		return 0, nil, nil
-	})
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			data = strings.TrimSpace(data)
-			if !strings.HasPrefix(data, "\"text\": \"") {
-				continue
-			}
-			data = strings.TrimPrefix(data, "\"text\": \"")
-			data = strings.TrimSuffix(data, "\"")
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+	scanner.Split(bufio.ScanLines)
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			// this is used to prevent annoying \ related format bug
-			data = fmt.Sprintf("{\"content\": \"%s\"}", data)
-			type dummyStruct struct {
-				Content string `json:"content"`
-			}
-			var dummy dummyStruct
-			err := json.Unmarshal([]byte(data), &dummy)
-			responseText += dummy.Content
-			var choice openai.ChatCompletionsStreamResponseChoice
-			choice.Delta.Content = dummy.Content
-			response := openai.ChatCompletionsStreamResponse{
-				Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
-				Object:  "chat.completion.chunk",
-				Created: helper.GetTimestamp(),
-				Model:   "gemini-pro",
-				Choices: []openai.ChatCompletionsStreamResponseChoice{choice},
-			}
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		data = strings.TrimSpace(data)
+		if !strings.HasPrefix(data, "data: ") {
+			continue
 		}
-	})
+		data = strings.TrimPrefix(data, "data: ")
+		data = strings.TrimSuffix(data, "\"")
+
+		var geminiResponse ChatResponse
+		err := json.Unmarshal([]byte(data), &geminiResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		response := streamResponseGeminiChat2OpenAI(&geminiResponse)
+		if response == nil {
+			continue
+		}
+
+		responseText += response.Choices[0].Delta.StringContent()
+
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), ""
 	}
+
 	return nil, responseText
 }
 
@@ -301,3 +392,39 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName st
 	_, err = c.Writer.Write(jsonResponse)
 	return nil, &usage
 }
+
+func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var geminiEmbeddingResponse EmbeddingResponse
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &geminiEmbeddingResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	if geminiEmbeddingResponse.Error != nil {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: geminiEmbeddingResponse.Error.Message,
+				Type:    "gemini_error",
+				Param:   "",
+				Code:    geminiEmbeddingResponse.Error.Code,
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+	fullTextResponse := embeddingResponseGemini2OpenAI(&geminiEmbeddingResponse)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &fullTextResponse.Usage
+}
diff --git a/relay/adaptor/gemini/model.go b/relay/adaptor/gemini/model.go
new file mode 100644
index 00000000..720cb65d
--- /dev/null
+++ b/relay/adaptor/gemini/model.go
@@ -0,0 +1,76 @@
+package gemini
+
+type ChatRequest struct {
+	Contents         []ChatContent        `json:"contents"`
+	SafetySettings   []ChatSafetySettings `json:"safety_settings,omitempty"`
+	GenerationConfig ChatGenerationConfig `json:"generation_config,omitempty"`
+	Tools            []ChatTools          `json:"tools,omitempty"`
+}
+
+type EmbeddingRequest struct {
+	Model                string      `json:"model"`
+	Content              ChatContent `json:"content"`
+	TaskType             string      `json:"taskType,omitempty"`
+	Title                string      `json:"title,omitempty"`
+	OutputDimensionality int         `json:"outputDimensionality,omitempty"`
+}
+
+type BatchEmbeddingRequest struct {
+	Requests []EmbeddingRequest `json:"requests"`
+}
+
+type EmbeddingData struct {
+	Values []float64 `json:"values"`
+}
+
+type EmbeddingResponse struct {
+	Embeddings []EmbeddingData `json:"embeddings"`
+	Error      *Error          `json:"error,omitempty"`
+}
+
+type Error struct {
+	Code    int    `json:"code,omitempty"`
+	Message string `json:"message,omitempty"`
+	Status  string `json:"status,omitempty"`
+}
+
+type InlineData struct {
+	MimeType string `json:"mimeType"`
+	Data     string `json:"data"`
+}
+
+type FunctionCall struct {
+	FunctionName string `json:"name"`
+	Arguments    any    `json:"args"`
+}
+
+type Part struct {
+	Text         string        `json:"text,omitempty"`
+	InlineData   *InlineData   `json:"inlineData,omitempty"`
+	FunctionCall *FunctionCall `json:"functionCall,omitempty"`
+}
+
+type ChatContent struct {
+	Role  string `json:"role,omitempty"`
+	Parts []Part `json:"parts"`
+}
+
+type ChatSafetySettings struct {
+	Category  string `json:"category"`
+	Threshold string `json:"threshold"`
+}
+
+type ChatTools struct {
+	FunctionDeclarations any `json:"function_declarations,omitempty"`
+}
+
+type ChatGenerationConfig struct {
+	ResponseMimeType string   `json:"responseMimeType,omitempty"`
+	ResponseSchema   any      `json:"responseSchema,omitempty"`
+	Temperature      *float64 `json:"temperature,omitempty"`
+	TopP             *float64 `json:"topP,omitempty"`
+	TopK             float64  `json:"topK,omitempty"`
+	MaxOutputTokens  int      `json:"maxOutputTokens,omitempty"`
+	CandidateCount   int      `json:"candidateCount,omitempty"`
+	StopSequences    []string `json:"stopSequences,omitempty"`
+}
diff --git a/relay/adaptor/groq/constants.go b/relay/adaptor/groq/constants.go
new file mode 100644
index 00000000..0864ebe7
--- /dev/null
+++ b/relay/adaptor/groq/constants.go
@@ -0,0 +1,27 @@
+package groq
+
+// https://console.groq.com/docs/models
+
+var ModelList = []string{
+	"gemma-7b-it",
+	"gemma2-9b-it",
+	"llama-3.1-70b-versatile",
+	"llama-3.1-8b-instant",
+	"llama-3.2-11b-text-preview",
+	"llama-3.2-11b-vision-preview",
+	"llama-3.2-1b-preview",
+	"llama-3.2-3b-preview",
+	"llama-3.2-11b-vision-preview",
+	"llama-3.2-90b-text-preview",
+	"llama-3.2-90b-vision-preview",
+	"llama-guard-3-8b",
+	"llama3-70b-8192",
+	"llama3-8b-8192",
+	"llama3-groq-70b-8192-tool-use-preview",
+	"llama3-groq-8b-8192-tool-use-preview",
+	"llava-v1.5-7b-4096-preview",
+	"mixtral-8x7b-32768",
+	"distil-whisper-large-v3-en",
+	"whisper-large-v3",
+	"whisper-large-v3-turbo",
+}
diff --git a/relay/adaptor/interface.go b/relay/adaptor/interface.go
new file mode 100644
index 00000000..01b2e2cb
--- /dev/null
+++ b/relay/adaptor/interface.go
@@ -0,0 +1,21 @@
+package adaptor
+
+import (
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+)
+
+type Adaptor interface {
+	Init(meta *meta.Meta)
+	GetRequestURL(meta *meta.Meta) (string, error)
+	SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	ConvertImageRequest(request *model.ImageRequest) (any, error)
+	DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error)
+	DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+	GetModelList() []string
+	GetChannelName() string
+}
diff --git a/relay/channel/lingyiwanwu/constants.go b/relay/adaptor/lingyiwanwu/constants.go
similarity index 100%
rename from relay/channel/lingyiwanwu/constants.go
rename to relay/adaptor/lingyiwanwu/constants.go
diff --git a/relay/adaptor/minimax/constants.go b/relay/adaptor/minimax/constants.go
new file mode 100644
index 00000000..1b2fc104
--- /dev/null
+++ b/relay/adaptor/minimax/constants.go
@@ -0,0 +1,11 @@
+package minimax
+
+// https://www.minimaxi.com/document/guides/chat-model/V2?id=65e0736ab2845de20908e2dd
+
+var ModelList = []string{
+	"abab6.5-chat",
+	"abab6.5s-chat",
+	"abab6-chat",
+	"abab5.5-chat",
+	"abab5.5s-chat",
+}
diff --git a/relay/adaptor/minimax/main.go b/relay/adaptor/minimax/main.go
new file mode 100644
index 00000000..fc9b5d26
--- /dev/null
+++ b/relay/adaptor/minimax/main.go
@@ -0,0 +1,14 @@
+package minimax
+
+import (
+	"fmt"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+func GetRequestURL(meta *meta.Meta) (string, error) {
+	if meta.Mode == relaymode.ChatCompletions {
+		return fmt.Sprintf("%s/v1/text/chatcompletion_v2", meta.BaseURL), nil
+	}
+	return "", fmt.Errorf("unsupported relay mode %d for minimax", meta.Mode)
+}
diff --git a/relay/channel/mistral/constants.go b/relay/adaptor/mistral/constants.go
similarity index 100%
rename from relay/channel/mistral/constants.go
rename to relay/adaptor/mistral/constants.go
diff --git a/relay/channel/moonshot/constants.go b/relay/adaptor/moonshot/constants.go
similarity index 100%
rename from relay/channel/moonshot/constants.go
rename to relay/adaptor/moonshot/constants.go
diff --git a/relay/adaptor/novita/constants.go b/relay/adaptor/novita/constants.go
new file mode 100644
index 00000000..c6618308
--- /dev/null
+++ b/relay/adaptor/novita/constants.go
@@ -0,0 +1,19 @@
+package novita
+
+// https://novita.ai/llm-api
+
+var ModelList = []string{
+	"meta-llama/llama-3-8b-instruct",
+	"meta-llama/llama-3-70b-instruct",
+	"nousresearch/hermes-2-pro-llama-3-8b",
+	"nousresearch/nous-hermes-llama2-13b",
+	"mistralai/mistral-7b-instruct",
+	"cognitivecomputations/dolphin-mixtral-8x22b",
+	"sao10k/l3-70b-euryale-v2.1",
+	"sophosympatheia/midnight-rose-70b",
+	"gryphe/mythomax-l2-13b",
+	"Nous-Hermes-2-Mixtral-8x7B-DPO",
+	"lzlv_70b",
+	"teknium/openhermes-2.5-mistral-7b",
+	"microsoft/wizardlm-2-8x22b",
+}
diff --git a/relay/adaptor/novita/main.go b/relay/adaptor/novita/main.go
new file mode 100644
index 00000000..80efa412
--- /dev/null
+++ b/relay/adaptor/novita/main.go
@@ -0,0 +1,15 @@
+package novita
+
+import (
+	"fmt"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+func GetRequestURL(meta *meta.Meta) (string, error) {
+	if meta.Mode == relaymode.ChatCompletions {
+		return fmt.Sprintf("%s/chat/completions", meta.BaseURL), nil
+	}
+	return "", fmt.Errorf("unsupported relay mode %d for novita", meta.Mode)
+}
diff --git a/relay/adaptor/ollama/adaptor.go b/relay/adaptor/ollama/adaptor.go
new file mode 100644
index 00000000..ad1f8983
--- /dev/null
+++ b/relay/adaptor/ollama/adaptor.go
@@ -0,0 +1,82 @@
+package ollama
+
+import (
+	"errors"
+	"fmt"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+	"io"
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	// https://github.com/ollama/ollama/blob/main/docs/api.md
+	fullRequestURL := fmt.Sprintf("%s/api/chat", meta.BaseURL)
+	if meta.Mode == relaymode.Embeddings {
+		fullRequestURL = fmt.Sprintf("%s/api/embed", meta.BaseURL)
+	}
+	return fullRequestURL, nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	switch relayMode {
+	case relaymode.Embeddings:
+		ollamaEmbeddingRequest := ConvertEmbeddingRequest(*request)
+		return ollamaEmbeddingRequest, nil
+	default:
+		return ConvertRequest(*request), nil
+	}
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		switch meta.Mode {
+		case relaymode.Embeddings:
+			err, usage = EmbeddingHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "ollama"
+}
diff --git a/relay/adaptor/ollama/constants.go b/relay/adaptor/ollama/constants.go
new file mode 100644
index 00000000..d9dc72a8
--- /dev/null
+++ b/relay/adaptor/ollama/constants.go
@@ -0,0 +1,11 @@
+package ollama
+
+var ModelList = []string{
+	"codellama:7b-instruct",
+	"llama2:7b",
+	"llama2:latest",
+	"llama3:latest",
+	"phi3:latest",
+	"qwen:0.5b-chat",
+	"qwen:7b",
+}
diff --git a/relay/channel/ollama/main.go b/relay/adaptor/ollama/main.go
similarity index 52%
rename from relay/channel/ollama/main.go
rename to relay/adaptor/ollama/main.go
index 7ec646a3..fa1b05f0 100644
--- a/relay/channel/ollama/main.go
+++ b/relay/adaptor/ollama/main.go
@@ -5,16 +5,21 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
+
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/random"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/image"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
@@ -26,13 +31,28 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 			TopP:             request.TopP,
 			FrequencyPenalty: request.FrequencyPenalty,
 			PresencePenalty:  request.PresencePenalty,
+			NumPredict:       request.MaxTokens,
+			NumCtx:           request.NumCtx,
 		},
 		Stream: request.Stream,
 	}
 	for _, message := range request.Messages {
+		openaiContent := message.ParseContent()
+		var imageUrls []string
+		var contentText string
+		for _, part := range openaiContent {
+			switch part.Type {
+			case model.ContentTypeText:
+				contentText = part.Text
+			case model.ContentTypeImageURL:
+				_, data, _ := image.GetImageFromUrl(part.ImageURL.Url)
+				imageUrls = append(imageUrls, data)
+			}
+		}
 		ollamaRequest.Messages = append(ollamaRequest.Messages, Message{
 			Role:    message.Role,
-			Content: message.StringContent(),
+			Content: contentText,
+			Images:  imageUrls,
 		})
 	}
 	return &ollamaRequest
@@ -50,7 +70,8 @@ func responseOllama2OpenAI(response *ChatResponse) *openai.TextResponse {
 		choice.FinishReason = "stop"
 	}
 	fullTextResponse := openai.TextResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
+		Model:   response.Model,
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
@@ -71,7 +92,7 @@ func streamResponseOllama2OpenAI(ollamaResponse *ChatResponse) *openai.ChatCompl
 		choice.FinishReason = &constant.StopFinishReason
 	}
 	response := openai.ChatCompletionsStreamResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   ollamaResponse.Model,
@@ -88,57 +109,123 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 			return 0, nil, nil
 		}
 		if i := strings.Index(string(data), "}\n"); i >= 0 {
-			return i + 2, data[0:i], nil
+			return i + 2, data[0 : i+1], nil
 		}
 		if atEOF {
 			return len(data), data, nil
 		}
 		return 0, nil, nil
 	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := strings.TrimPrefix(scanner.Text(), "}")
-			dataChan <- data + "}"
-		}
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			var ollamaResponse ChatResponse
-			err := json.Unmarshal([]byte(data), &ollamaResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			if ollamaResponse.EvalCount != 0 {
-				usage.PromptTokens = ollamaResponse.PromptEvalCount
-				usage.CompletionTokens = ollamaResponse.EvalCount
-				usage.TotalTokens = ollamaResponse.PromptEvalCount + ollamaResponse.EvalCount
-			}
-			response := streamResponseOllama2OpenAI(&ollamaResponse)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if strings.HasPrefix(data, "}") {
+			data = strings.TrimPrefix(data, "}") + "}"
 		}
-	})
+
+		var ollamaResponse ChatResponse
+		err := json.Unmarshal([]byte(data), &ollamaResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		if ollamaResponse.EvalCount != 0 {
+			usage.PromptTokens = ollamaResponse.PromptEvalCount
+			usage.CompletionTokens = ollamaResponse.EvalCount
+			usage.TotalTokens = ollamaResponse.PromptEvalCount + ollamaResponse.EvalCount
+		}
+
+		response := streamResponseOllama2OpenAI(&ollamaResponse)
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
+
 	return nil, &usage
 }
 
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
+	return &EmbeddingRequest{
+		Model: request.Model,
+		Input: request.ParseInput(),
+		Options: &Options{
+			Seed:             int(request.Seed),
+			Temperature:      request.Temperature,
+			TopP:             request.TopP,
+			FrequencyPenalty: request.FrequencyPenalty,
+			PresencePenalty:  request.PresencePenalty,
+		},
+	}
+}
+
+func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var ollamaResponse EmbeddingResponse
+	err := json.NewDecoder(resp.Body).Decode(&ollamaResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	if ollamaResponse.Error != "" {
+		return &model.ErrorWithStatusCode{
+			Error: model.Error{
+				Message: ollamaResponse.Error,
+				Type:    "ollama_error",
+				Param:   "",
+				Code:    "ollama_error",
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+
+	fullTextResponse := embeddingResponseOllama2OpenAI(&ollamaResponse)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &fullTextResponse.Usage
+}
+
+func embeddingResponseOllama2OpenAI(response *EmbeddingResponse) *openai.EmbeddingResponse {
+	openAIEmbeddingResponse := openai.EmbeddingResponse{
+		Object: "list",
+		Data:   make([]openai.EmbeddingResponseItem, 0, 1),
+		Model:  response.Model,
+		Usage:  model.Usage{TotalTokens: 0},
+	}
+
+	for i, embedding := range response.Embeddings {
+		openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
+			Object:    `embedding`,
+			Index:     i,
+			Embedding: embedding,
+		})
+	}
+	return &openAIEmbeddingResponse
+}
+
 func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	ctx := context.TODO()
 	var ollamaResponse ChatResponse
diff --git a/relay/channel/ollama/model.go b/relay/adaptor/ollama/model.go
similarity index 54%
rename from relay/channel/ollama/model.go
rename to relay/adaptor/ollama/model.go
index a8ef1ffc..94f2ab73 100644
--- a/relay/channel/ollama/model.go
+++ b/relay/adaptor/ollama/model.go
@@ -1,12 +1,14 @@
 package ollama
 
 type Options struct {
-	Seed             int     `json:"seed,omitempty"`
-	Temperature      float64 `json:"temperature,omitempty"`
-	TopK             int     `json:"top_k,omitempty"`
-	TopP             float64 `json:"top_p,omitempty"`
-	FrequencyPenalty float64 `json:"frequency_penalty,omitempty"`
-	PresencePenalty  float64 `json:"presence_penalty,omitempty"`
+	Seed             int      `json:"seed,omitempty"`
+	Temperature      *float64 `json:"temperature,omitempty"`
+	TopK             int      `json:"top_k,omitempty"`
+	TopP             *float64 `json:"top_p,omitempty"`
+	FrequencyPenalty *float64 `json:"frequency_penalty,omitempty"`
+	PresencePenalty  *float64 `json:"presence_penalty,omitempty"`
+	NumPredict       int      `json:"num_predict,omitempty"`
+	NumCtx           int      `json:"num_ctx,omitempty"`
 }
 
 type Message struct {
@@ -35,3 +37,17 @@ type ChatResponse struct {
 	EvalDuration    int     `json:"eval_duration,omitempty"`
 	Error           string  `json:"error,omitempty"`
 }
+
+type EmbeddingRequest struct {
+	Model string   `json:"model"`
+	Input []string `json:"input"`
+	// Truncate  bool     `json:"truncate,omitempty"`
+	Options *Options `json:"options,omitempty"`
+	// KeepAlive string   `json:"keep_alive,omitempty"`
+}
+
+type EmbeddingResponse struct {
+	Error      string      `json:"error,omitempty"`
+	Model      string      `json:"model"`
+	Embeddings [][]float64 `json:"embeddings"`
+}
diff --git a/relay/adaptor/openai/adaptor.go b/relay/adaptor/openai/adaptor.go
new file mode 100644
index 00000000..6946e402
--- /dev/null
+++ b/relay/adaptor/openai/adaptor.go
@@ -0,0 +1,129 @@
+package openai
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/doubao"
+	"github.com/songquanpeng/one-api/relay/adaptor/minimax"
+	"github.com/songquanpeng/one-api/relay/adaptor/novita"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+type Adaptor struct {
+	ChannelType int
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.ChannelType = meta.ChannelType
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	switch meta.ChannelType {
+	case channeltype.Azure:
+		if meta.Mode == relaymode.ImagesGenerations {
+			// https://learn.microsoft.com/en-us/azure/ai-services/openai/dall-e-quickstart?tabs=dalle3%2Ccommand-line&pivots=rest-api
+			// https://{resource_name}.openai.azure.com/openai/deployments/dall-e-3/images/generations?api-version=2024-03-01-preview
+			fullRequestURL := fmt.Sprintf("%s/openai/deployments/%s/images/generations?api-version=%s", meta.BaseURL, meta.ActualModelName, meta.Config.APIVersion)
+			return fullRequestURL, nil
+		}
+
+		// https://learn.microsoft.com/en-us/azure/cognitive-services/openai/chatgpt-quickstart?pivots=rest-api&tabs=command-line#rest-api
+		requestURL := strings.Split(meta.RequestURLPath, "?")[0]
+		requestURL = fmt.Sprintf("%s?api-version=%s", requestURL, meta.Config.APIVersion)
+		task := strings.TrimPrefix(requestURL, "/v1/")
+		model_ := meta.ActualModelName
+		model_ = strings.Replace(model_, ".", "", -1)
+		//https://github.com/songquanpeng/one-api/issues/1191
+		// {your endpoint}/openai/deployments/{your azure_model}/chat/completions?api-version={api_version}
+		requestURL = fmt.Sprintf("/openai/deployments/%s/%s", model_, task)
+		return GetFullRequestURL(meta.BaseURL, requestURL, meta.ChannelType), nil
+	case channeltype.Minimax:
+		return minimax.GetRequestURL(meta)
+	case channeltype.Doubao:
+		return doubao.GetRequestURL(meta)
+	case channeltype.Novita:
+		return novita.GetRequestURL(meta)
+	default:
+		return GetFullRequestURL(meta.BaseURL, meta.RequestURLPath, meta.ChannelType), nil
+	}
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	if meta.ChannelType == channeltype.Azure {
+		req.Header.Set("api-key", meta.APIKey)
+		return nil
+	}
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	if meta.ChannelType == channeltype.OpenRouter {
+		req.Header.Set("HTTP-Referer", "https://github.com/songquanpeng/one-api")
+		req.Header.Set("X-Title", "One API")
+	}
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	if request.Stream {
+		// always return usage in stream mode
+		if request.StreamOptions == nil {
+			request.StreamOptions = &model.StreamOptions{}
+		}
+		request.StreamOptions.IncludeUsage = true
+	}
+	return request, nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText, usage = StreamHandler(c, resp, meta.Mode)
+		if usage == nil || usage.TotalTokens == 0 {
+			usage = ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+		}
+		if usage.TotalTokens != 0 && usage.PromptTokens == 0 { // some channels don't return prompt tokens & completion tokens
+			usage.PromptTokens = meta.PromptTokens
+			usage.CompletionTokens = usage.TotalTokens - meta.PromptTokens
+		}
+	} else {
+		switch meta.Mode {
+		case relaymode.ImagesGenerations:
+			err, _ = ImageHandler(c, resp)
+		default:
+			err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+		}
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	_, modelList := GetCompatibleChannelMeta(a.ChannelType)
+	return modelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	channelName, _ := GetCompatibleChannelMeta(a.ChannelType)
+	return channelName
+}
diff --git a/relay/adaptor/openai/compatible.go b/relay/adaptor/openai/compatible.go
new file mode 100644
index 00000000..15b4dcc0
--- /dev/null
+++ b/relay/adaptor/openai/compatible.go
@@ -0,0 +1,74 @@
+package openai
+
+import (
+	"github.com/songquanpeng/one-api/relay/adaptor/ai360"
+	"github.com/songquanpeng/one-api/relay/adaptor/baichuan"
+	"github.com/songquanpeng/one-api/relay/adaptor/deepseek"
+	"github.com/songquanpeng/one-api/relay/adaptor/doubao"
+	"github.com/songquanpeng/one-api/relay/adaptor/groq"
+	"github.com/songquanpeng/one-api/relay/adaptor/lingyiwanwu"
+	"github.com/songquanpeng/one-api/relay/adaptor/minimax"
+	"github.com/songquanpeng/one-api/relay/adaptor/mistral"
+	"github.com/songquanpeng/one-api/relay/adaptor/moonshot"
+	"github.com/songquanpeng/one-api/relay/adaptor/novita"
+	"github.com/songquanpeng/one-api/relay/adaptor/siliconflow"
+	"github.com/songquanpeng/one-api/relay/adaptor/stepfun"
+	"github.com/songquanpeng/one-api/relay/adaptor/togetherai"
+	"github.com/songquanpeng/one-api/relay/adaptor/xai"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+)
+
+var CompatibleChannels = []int{
+	channeltype.Azure,
+	channeltype.AI360,
+	channeltype.Moonshot,
+	channeltype.Baichuan,
+	channeltype.Minimax,
+	channeltype.Doubao,
+	channeltype.Mistral,
+	channeltype.Groq,
+	channeltype.LingYiWanWu,
+	channeltype.StepFun,
+	channeltype.DeepSeek,
+	channeltype.TogetherAI,
+	channeltype.Novita,
+	channeltype.SiliconFlow,
+	channeltype.XAI,
+}
+
+func GetCompatibleChannelMeta(channelType int) (string, []string) {
+	switch channelType {
+	case channeltype.Azure:
+		return "azure", ModelList
+	case channeltype.AI360:
+		return "360", ai360.ModelList
+	case channeltype.Moonshot:
+		return "moonshot", moonshot.ModelList
+	case channeltype.Baichuan:
+		return "baichuan", baichuan.ModelList
+	case channeltype.Minimax:
+		return "minimax", minimax.ModelList
+	case channeltype.Mistral:
+		return "mistralai", mistral.ModelList
+	case channeltype.Groq:
+		return "groq", groq.ModelList
+	case channeltype.LingYiWanWu:
+		return "lingyiwanwu", lingyiwanwu.ModelList
+	case channeltype.StepFun:
+		return "stepfun", stepfun.ModelList
+	case channeltype.DeepSeek:
+		return "deepseek", deepseek.ModelList
+	case channeltype.TogetherAI:
+		return "together.ai", togetherai.ModelList
+	case channeltype.Doubao:
+		return "doubao", doubao.ModelList
+	case channeltype.Novita:
+		return "novita", novita.ModelList
+	case channeltype.SiliconFlow:
+		return "siliconflow", siliconflow.ModelList
+	case channeltype.XAI:
+		return "xai", xai.ModelList
+	default:
+		return "openai", ModelList
+	}
+}
diff --git a/relay/channel/openai/constants.go b/relay/adaptor/openai/constants.go
similarity index 71%
rename from relay/channel/openai/constants.go
rename to relay/adaptor/openai/constants.go
index ea236ea1..8a643bc6 100644
--- a/relay/channel/openai/constants.go
+++ b/relay/adaptor/openai/constants.go
@@ -6,7 +6,12 @@ var ModelList = []string{
 	"gpt-3.5-turbo-instruct",
 	"gpt-4", "gpt-4-0314", "gpt-4-0613", "gpt-4-1106-preview", "gpt-4-0125-preview",
 	"gpt-4-32k", "gpt-4-32k-0314", "gpt-4-32k-0613",
-	"gpt-4-turbo-preview",
+	"gpt-4-turbo-preview", "gpt-4-turbo", "gpt-4-turbo-2024-04-09",
+	"gpt-4o", "gpt-4o-2024-05-13",
+	"gpt-4o-2024-08-06",
+	"gpt-4o-2024-11-20",
+	"chatgpt-4o-latest",
+	"gpt-4o-mini", "gpt-4o-mini-2024-07-18",
 	"gpt-4-vision-preview",
 	"text-embedding-ada-002", "text-embedding-3-small", "text-embedding-3-large",
 	"text-curie-001", "text-babbage-001", "text-ada-001", "text-davinci-002", "text-davinci-003",
@@ -16,4 +21,7 @@ var ModelList = []string{
 	"dall-e-2", "dall-e-3",
 	"whisper-1",
 	"tts-1", "tts-1-1106", "tts-1-hd", "tts-1-hd-1106",
+	"o1", "o1-2024-12-17",
+	"o1-preview", "o1-preview-2024-09-12",
+	"o1-mini", "o1-mini-2024-09-12",
 }
diff --git a/relay/adaptor/openai/helper.go b/relay/adaptor/openai/helper.go
new file mode 100644
index 00000000..47c2a882
--- /dev/null
+++ b/relay/adaptor/openai/helper.go
@@ -0,0 +1,31 @@
+package openai
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+func ResponseText2Usage(responseText string, modelName string, promptTokens int) *model.Usage {
+	usage := &model.Usage{}
+	usage.PromptTokens = promptTokens
+	usage.CompletionTokens = CountTokenText(responseText, modelName)
+	usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
+	return usage
+}
+
+func GetFullRequestURL(baseURL string, requestURL string, channelType int) string {
+	fullRequestURL := fmt.Sprintf("%s%s", baseURL, requestURL)
+
+	if strings.HasPrefix(baseURL, "https://gateway.ai.cloudflare.com") {
+		switch channelType {
+		case channeltype.OpenAI:
+			fullRequestURL = fmt.Sprintf("%s%s", baseURL, strings.TrimPrefix(requestURL, "/v1"))
+		case channeltype.Azure:
+			fullRequestURL = fmt.Sprintf("%s%s", baseURL, strings.TrimPrefix(requestURL, "/openai/deployments"))
+		}
+	}
+	return fullRequestURL
+}
diff --git a/relay/adaptor/openai/image.go b/relay/adaptor/openai/image.go
new file mode 100644
index 00000000..0f89618a
--- /dev/null
+++ b/relay/adaptor/openai/image.go
@@ -0,0 +1,44 @@
+package openai
+
+import (
+	"bytes"
+	"encoding/json"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+)
+
+func ImageHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var imageResponse ImageResponse
+	responseBody, err := io.ReadAll(resp.Body)
+
+	if err != nil {
+		return ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &imageResponse)
+	if err != nil {
+		return ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	resp.Body = io.NopCloser(bytes.NewBuffer(responseBody))
+
+	for k, v := range resp.Header {
+		c.Writer.Header().Set(k, v[0])
+	}
+	c.Writer.WriteHeader(resp.StatusCode)
+
+	_, err = io.Copy(c.Writer, resp.Body)
+	if err != nil {
+		return ErrorWrapper(err, "copy_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	return nil, nil
+}
diff --git a/relay/channel/openai/main.go b/relay/adaptor/openai/main.go
similarity index 56%
rename from relay/channel/openai/main.go
rename to relay/adaptor/openai/main.go
index d47cd164..97080738 100644
--- a/relay/channel/openai/main.go
+++ b/relay/adaptor/openai/main.go
@@ -4,94 +4,95 @@ import (
 	"bufio"
 	"bytes"
 	"encoding/json"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
 	"strings"
+
+	"github.com/songquanpeng/one-api/common/render"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/conv"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+const (
+	dataPrefix       = "data: "
+	done             = "[DONE]"
+	dataPrefixLength = len(dataPrefix)
 )
 
 func StreamHandler(c *gin.Context, resp *http.Response, relayMode int) (*model.ErrorWithStatusCode, string, *model.Usage) {
 	responseText := ""
 	scanner := bufio.NewScanner(resp.Body)
-	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
-		if atEOF && len(data) == 0 {
-			return 0, nil, nil
-		}
-		if i := strings.Index(string(data), "\n"); i >= 0 {
-			return i + 1, data[0:i], nil
-		}
-		if atEOF {
-			return len(data), data, nil
-		}
-		return 0, nil, nil
-	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
+	scanner.Split(bufio.ScanLines)
 	var usage *model.Usage
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 6 { // ignore blank line or wrong format
-				continue
-			}
-			if data[:6] != "data: " && data[:6] != "[DONE]" {
-				continue
-			}
-			dataChan <- data
-			data = data[6:]
-			if !strings.HasPrefix(data, "[DONE]") {
-				switch relayMode {
-				case constant.RelayModeChatCompletions:
-					var streamResponse ChatCompletionsStreamResponse
-					err := json.Unmarshal([]byte(data), &streamResponse)
-					if err != nil {
-						logger.SysError("error unmarshalling stream response: " + err.Error())
-						continue // just ignore the error
-					}
-					for _, choice := range streamResponse.Choices {
-						responseText += choice.Delta.Content
-					}
-					if streamResponse.Usage != nil {
-						usage = streamResponse.Usage
-					}
-				case constant.RelayModeCompletions:
-					var streamResponse CompletionsStreamResponse
-					err := json.Unmarshal([]byte(data), &streamResponse)
-					if err != nil {
-						logger.SysError("error unmarshalling stream response: " + err.Error())
-						continue
-					}
-					for _, choice := range streamResponse.Choices {
-						responseText += choice.Text
-					}
-				}
-			}
-		}
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			if strings.HasPrefix(data, "data: [DONE]") {
-				data = data[:12]
-			}
-			// some implementations may add \r at the end of data
-			data = strings.TrimSuffix(data, "\r")
-			c.Render(-1, common.CustomEvent{Data: data})
-			return true
-		case <-stopChan:
-			return false
+
+	doneRendered := false
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < dataPrefixLength { // ignore blank line or wrong format
+			continue
 		}
-	})
+		if data[:dataPrefixLength] != dataPrefix && data[:dataPrefixLength] != done {
+			continue
+		}
+		if strings.HasPrefix(data[dataPrefixLength:], done) {
+			render.StringData(c, data)
+			doneRendered = true
+			continue
+		}
+		switch relayMode {
+		case relaymode.ChatCompletions:
+			var streamResponse ChatCompletionsStreamResponse
+			err := json.Unmarshal([]byte(data[dataPrefixLength:]), &streamResponse)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				render.StringData(c, data) // if error happened, pass the data to client
+				continue                   // just ignore the error
+			}
+			if len(streamResponse.Choices) == 0 && streamResponse.Usage == nil {
+				// but for empty choice and no usage, we should not pass it to client, this is for azure
+				continue // just ignore empty choice
+			}
+			render.StringData(c, data)
+			for _, choice := range streamResponse.Choices {
+				responseText += conv.AsString(choice.Delta.Content)
+			}
+			if streamResponse.Usage != nil {
+				usage = streamResponse.Usage
+			}
+		case relaymode.Completions:
+			render.StringData(c, data)
+			var streamResponse CompletionsStreamResponse
+			err := json.Unmarshal([]byte(data[dataPrefixLength:]), &streamResponse)
+			if err != nil {
+				logger.SysError("error unmarshalling stream response: " + err.Error())
+				continue
+			}
+			for _, choice := range streamResponse.Choices {
+				responseText += choice.Text
+			}
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	if !doneRendered {
+		render.Done(c)
+	}
+
 	err := resp.Body.Close()
 	if err != nil {
 		return ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), "", nil
 	}
+
 	return nil, responseText, usage
 }
 
@@ -135,7 +136,7 @@ func Handler(c *gin.Context, resp *http.Response, promptTokens int, modelName st
 		return ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
 
-	if textResponse.Usage.TotalTokens == 0 {
+	if textResponse.Usage.TotalTokens == 0 || (textResponse.Usage.PromptTokens == 0 && textResponse.Usage.CompletionTokens == 0) {
 		completionTokens := 0
 		for _, choice := range textResponse.Choices {
 			completionTokens += CountTokenText(choice.Message.StringContent(), modelName)
diff --git a/relay/channel/openai/model.go b/relay/adaptor/openai/model.go
similarity index 88%
rename from relay/channel/openai/model.go
rename to relay/adaptor/openai/model.go
index 6c0b2c53..4c974de4 100644
--- a/relay/channel/openai/model.go
+++ b/relay/adaptor/openai/model.go
@@ -110,20 +110,22 @@ type EmbeddingResponse struct {
 	model.Usage `json:"usage"`
 }
 
+type ImageData struct {
+	Url           string `json:"url,omitempty"`
+	B64Json       string `json:"b64_json,omitempty"`
+	RevisedPrompt string `json:"revised_prompt,omitempty"`
+}
+
 type ImageResponse struct {
-	Created int `json:"created"`
-	Data    []struct {
-		Url string `json:"url"`
-	}
+	Created int64       `json:"created"`
+	Data    []ImageData `json:"data"`
+	//model.Usage `json:"usage"`
 }
 
 type ChatCompletionsStreamResponseChoice struct {
-	Index int `json:"index"`
-	Delta struct {
-		Content string `json:"content"`
-		Role    string `json:"role,omitempty"`
-	} `json:"delta"`
-	FinishReason *string `json:"finish_reason,omitempty"`
+	Index        int           `json:"index"`
+	Delta        model.Message `json:"delta"`
+	FinishReason *string       `json:"finish_reason,omitempty"`
 }
 
 type ChatCompletionsStreamResponse struct {
@@ -132,7 +134,7 @@ type ChatCompletionsStreamResponse struct {
 	Created int64                                 `json:"created"`
 	Model   string                                `json:"model"`
 	Choices []ChatCompletionsStreamResponseChoice `json:"choices"`
-	Usage   *model.Usage                          `json:"usage"`
+	Usage   *model.Usage                          `json:"usage,omitempty"`
 }
 
 type CompletionsStreamResponse struct {
diff --git a/relay/channel/openai/token.go b/relay/adaptor/openai/token.go
similarity index 84%
rename from relay/channel/openai/token.go
rename to relay/adaptor/openai/token.go
index 0720425f..7c8468b9 100644
--- a/relay/channel/openai/token.go
+++ b/relay/adaptor/openai/token.go
@@ -4,10 +4,10 @@ import (
 	"errors"
 	"fmt"
 	"github.com/pkoukk/tiktoken-go"
-	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/image"
 	"github.com/songquanpeng/one-api/common/logger"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
 	"github.com/songquanpeng/one-api/relay/model"
 	"math"
 	"strings"
@@ -24,13 +24,19 @@ func InitTokenEncoders() {
 		logger.FatalLog(fmt.Sprintf("failed to get gpt-3.5-turbo token encoder: %s", err.Error()))
 	}
 	defaultTokenEncoder = gpt35TokenEncoder
+	gpt4oTokenEncoder, err := tiktoken.EncodingForModel("gpt-4o")
+	if err != nil {
+		logger.FatalLog(fmt.Sprintf("failed to get gpt-4o token encoder: %s", err.Error()))
+	}
 	gpt4TokenEncoder, err := tiktoken.EncodingForModel("gpt-4")
 	if err != nil {
 		logger.FatalLog(fmt.Sprintf("failed to get gpt-4 token encoder: %s", err.Error()))
 	}
-	for model := range common.ModelRatio {
+	for model := range billingratio.ModelRatio {
 		if strings.HasPrefix(model, "gpt-3.5") {
 			tokenEncoderMap[model] = gpt35TokenEncoder
+		} else if strings.HasPrefix(model, "gpt-4o") {
+			tokenEncoderMap[model] = gpt4oTokenEncoder
 		} else if strings.HasPrefix(model, "gpt-4") {
 			tokenEncoderMap[model] = gpt4TokenEncoder
 		} else {
@@ -91,7 +97,11 @@ func CountTokenMessages(messages []model.Message, model string) int {
 				m := it.(map[string]any)
 				switch m["type"] {
 				case "text":
-					tokenNum += getTokenNum(tokenEncoder, m["text"].(string))
+					if textValue, ok := m["text"]; ok {
+						if textString, ok := textValue.(string); ok {
+							tokenNum += getTokenNum(tokenEncoder, textString)
+						}
+					}
 				case "image_url":
 					imageUrl, ok := m["image_url"].(map[string]any)
 					if ok {
@@ -100,7 +110,7 @@ func CountTokenMessages(messages []model.Message, model string) int {
 						if imageUrl["detail"] != nil {
 							detail = imageUrl["detail"].(string)
 						}
-						imageTokens, err := countImageTokens(url, detail)
+						imageTokens, err := countImageTokens(url, detail, model)
 						if err != nil {
 							logger.SysError("error counting image tokens: " + err.Error())
 						} else {
@@ -124,11 +134,15 @@ const (
 	lowDetailCost         = 85
 	highDetailCostPerTile = 170
 	additionalCost        = 85
+	// gpt-4o-mini cost higher than other model
+	gpt4oMiniLowDetailCost  = 2833
+	gpt4oMiniHighDetailCost = 5667
+	gpt4oMiniAdditionalCost = 2833
 )
 
 // https://platform.openai.com/docs/guides/vision/calculating-costs
 // https://github.com/openai/openai-cookbook/blob/05e3f9be4c7a2ae7ecf029a7c32065b024730ebe/examples/How_to_count_tokens_with_tiktoken.ipynb
-func countImageTokens(url string, detail string) (_ int, err error) {
+func countImageTokens(url string, detail string, model string) (_ int, err error) {
 	var fetchSize = true
 	var width, height int
 	// Reference: https://platform.openai.com/docs/guides/vision/low-or-high-fidelity-image-understanding
@@ -162,6 +176,9 @@ func countImageTokens(url string, detail string) (_ int, err error) {
 	}
 	switch detail {
 	case "low":
+		if strings.HasPrefix(model, "gpt-4o-mini") {
+			return gpt4oMiniLowDetailCost, nil
+		}
 		return lowDetailCost, nil
 	case "high":
 		if fetchSize {
@@ -181,6 +198,9 @@ func countImageTokens(url string, detail string) (_ int, err error) {
 			height = int(float64(height) * ratio)
 		}
 		numSquares := int(math.Ceil(float64(width)/512) * math.Ceil(float64(height)/512))
+		if strings.HasPrefix(model, "gpt-4o-mini") {
+			return numSquares*gpt4oMiniHighDetailCost + gpt4oMiniAdditionalCost, nil
+		}
 		result := numSquares*highDetailCostPerTile + additionalCost
 		return result, nil
 	default:
@@ -206,3 +226,7 @@ func CountTokenText(text string, model string) int {
 	tokenEncoder := getTokenEncoder(model)
 	return getTokenNum(tokenEncoder, text)
 }
+
+func CountToken(text string) int {
+	return CountTokenInput(text, "gpt-3.5-turbo")
+}
diff --git a/relay/channel/openai/util.go b/relay/adaptor/openai/util.go
similarity index 59%
rename from relay/channel/openai/util.go
rename to relay/adaptor/openai/util.go
index ba0cab7d..83beadba 100644
--- a/relay/channel/openai/util.go
+++ b/relay/adaptor/openai/util.go
@@ -1,8 +1,16 @@
 package openai
 
-import "github.com/songquanpeng/one-api/relay/model"
+import (
+	"context"
+	"fmt"
+
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/model"
+)
 
 func ErrorWrapper(err error, code string, statusCode int) *model.ErrorWithStatusCode {
+	logger.Error(context.TODO(), fmt.Sprintf("[%s]%+v", code, err))
+
 	Error := model.Error{
 		Message: err.Error(),
 		Type:    "one_api_error",
diff --git a/relay/channel/palm/adaptor.go b/relay/adaptor/palm/adaptor.go
similarity index 59%
rename from relay/channel/palm/adaptor.go
rename to relay/adaptor/palm/adaptor.go
index efd0620c..98aa3e18 100644
--- a/relay/channel/palm/adaptor.go
+++ b/relay/adaptor/palm/adaptor.go
@@ -4,10 +4,10 @@ import (
 	"errors"
 	"fmt"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 )
@@ -15,16 +15,16 @@ import (
 type Adaptor struct {
 }
 
-func (a *Adaptor) Init(meta *util.RelayMeta) {
+func (a *Adaptor) Init(meta *meta.Meta) {
 
 }
 
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	return fmt.Sprintf("%s/v1beta2/models/chat-bison-001:generateMessage", meta.BaseURL), nil
 }
 
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
 	req.Header.Set("x-goog-api-key", meta.APIKey)
 	return nil
 }
@@ -36,11 +36,18 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	return ConvertRequest(*request), nil
 }
 
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
 	if meta.IsStream {
 		var responseText string
 		err, responseText = StreamHandler(c, resp)
diff --git a/relay/channel/palm/constants.go b/relay/adaptor/palm/constants.go
similarity index 100%
rename from relay/channel/palm/constants.go
rename to relay/adaptor/palm/constants.go
diff --git a/relay/channel/palm/model.go b/relay/adaptor/palm/model.go
similarity index 72%
rename from relay/channel/palm/model.go
rename to relay/adaptor/palm/model.go
index f653022c..2bdd8f29 100644
--- a/relay/channel/palm/model.go
+++ b/relay/adaptor/palm/model.go
@@ -19,11 +19,11 @@ type Prompt struct {
 }
 
 type ChatRequest struct {
-	Prompt         Prompt  `json:"prompt"`
-	Temperature    float64 `json:"temperature,omitempty"`
-	CandidateCount int     `json:"candidateCount,omitempty"`
-	TopP           float64 `json:"topP,omitempty"`
-	TopK           int     `json:"topK,omitempty"`
+	Prompt         Prompt   `json:"prompt"`
+	Temperature    *float64 `json:"temperature,omitempty"`
+	CandidateCount int      `json:"candidateCount,omitempty"`
+	TopP           *float64 `json:"topP,omitempty"`
+	TopK           int      `json:"topK,omitempty"`
 }
 
 type Error struct {
diff --git a/relay/channel/palm/palm.go b/relay/adaptor/palm/palm.go
similarity index 74%
rename from relay/channel/palm/palm.go
rename to relay/adaptor/palm/palm.go
index 56738544..d31784ec 100644
--- a/relay/channel/palm/palm.go
+++ b/relay/adaptor/palm/palm.go
@@ -3,15 +3,18 @@ package palm
 import (
 	"encoding/json"
 	"fmt"
+	"github.com/songquanpeng/one-api/common/render"
+	"io"
+	"net/http"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/helper"
 	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
 )
 
 // https://developers.generativeai.google/api/rest/generativelanguage/models/generateMessage#request-body
@@ -74,60 +77,53 @@ func streamResponsePaLM2OpenAI(palmResponse *ChatResponse) *openai.ChatCompletio
 
 func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, string) {
 	responseText := ""
-	responseId := fmt.Sprintf("chatcmpl-%s", helper.GetUUID())
+	responseId := fmt.Sprintf("chatcmpl-%s", random.GetUUID())
 	createdTime := helper.GetTimestamp()
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		responseBody, err := io.ReadAll(resp.Body)
-		if err != nil {
-			logger.SysError("error reading stream response: " + err.Error())
-			stopChan <- true
-			return
-		}
-		err = resp.Body.Close()
-		if err != nil {
-			logger.SysError("error closing stream response: " + err.Error())
-			stopChan <- true
-			return
-		}
-		var palmResponse ChatResponse
-		err = json.Unmarshal(responseBody, &palmResponse)
-		if err != nil {
-			logger.SysError("error unmarshalling stream response: " + err.Error())
-			stopChan <- true
-			return
-		}
-		fullTextResponse := streamResponsePaLM2OpenAI(&palmResponse)
-		fullTextResponse.Id = responseId
-		fullTextResponse.Created = createdTime
-		if len(palmResponse.Candidates) > 0 {
-			responseText = palmResponse.Candidates[0].Content
-		}
-		jsonResponse, err := json.Marshal(fullTextResponse)
-		if err != nil {
-			logger.SysError("error marshalling stream response: " + err.Error())
-			stopChan <- true
-			return
-		}
-		dataChan <- string(jsonResponse)
-		stopChan <- true
-	}()
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			c.Render(-1, common.CustomEvent{Data: "data: " + data})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		logger.SysError("error reading stream response: " + err.Error())
+		err := resp.Body.Close()
+		if err != nil {
+			return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), ""
 		}
-	})
-	err := resp.Body.Close()
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), ""
+	}
+
+	err = resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), ""
 	}
+
+	var palmResponse ChatResponse
+	err = json.Unmarshal(responseBody, &palmResponse)
+	if err != nil {
+		logger.SysError("error unmarshalling stream response: " + err.Error())
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), ""
+	}
+
+	fullTextResponse := streamResponsePaLM2OpenAI(&palmResponse)
+	fullTextResponse.Id = responseId
+	fullTextResponse.Created = createdTime
+	if len(palmResponse.Candidates) > 0 {
+		responseText = palmResponse.Candidates[0].Content
+	}
+
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		logger.SysError("error marshalling stream response: " + err.Error())
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), ""
+	}
+
+	err = render.ObjectData(c, string(jsonResponse))
+	if err != nil {
+		logger.SysError(err.Error())
+	}
+
+	render.Done(c)
+
 	return nil, responseText
 }
 
diff --git a/relay/adaptor/proxy/adaptor.go b/relay/adaptor/proxy/adaptor.go
new file mode 100644
index 00000000..670c7628
--- /dev/null
+++ b/relay/adaptor/proxy/adaptor.go
@@ -0,0 +1,89 @@
+package proxy
+
+import (
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ adaptor.Adaptor = new(Adaptor)
+
+const channelName = "proxy"
+
+type Adaptor struct{}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	return nil, errors.New("notimplement")
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	for k, v := range resp.Header {
+		for _, vv := range v {
+			c.Writer.Header().Set(k, vv)
+		}
+	}
+
+	c.Writer.WriteHeader(resp.StatusCode)
+	if _, gerr := io.Copy(c.Writer, resp.Body); gerr != nil {
+		return nil, &relaymodel.ErrorWithStatusCode{
+			StatusCode: http.StatusInternalServerError,
+			Error: relaymodel.Error{
+				Message: gerr.Error(),
+			},
+		}
+	}
+
+	return nil, nil
+}
+
+func (a *Adaptor) GetModelList() (models []string) {
+	return nil
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return channelName
+}
+
+// GetRequestURL remove static prefix, and return the real request url to the upstream service
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	prefix := fmt.Sprintf("/v1/oneapi/proxy/%d", meta.ChannelId)
+	return meta.BaseURL + strings.TrimPrefix(meta.RequestURLPath, prefix), nil
+
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	for k, v := range c.Request.Header {
+		req.Header.Set(k, v[0])
+	}
+
+	// remove unnecessary headers
+	req.Header.Del("Host")
+	req.Header.Del("Content-Length")
+	req.Header.Del("Accept-Encoding")
+	req.Header.Del("Connection")
+
+	// set authorization header
+	req.Header.Set("Authorization", meta.APIKey)
+
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	return nil, errors.Errorf("not implement")
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
+}
diff --git a/relay/adaptor/replicate/adaptor.go b/relay/adaptor/replicate/adaptor.go
new file mode 100644
index 00000000..a60a7de3
--- /dev/null
+++ b/relay/adaptor/replicate/adaptor.go
@@ -0,0 +1,136 @@
+package replicate
+
+import (
+	"fmt"
+	"io"
+	"net/http"
+	"slices"
+	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+)
+
+type Adaptor struct {
+	meta *meta.Meta
+}
+
+// ConvertImageRequest implements adaptor.Adaptor.
+func (*Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	return DrawImageRequest{
+		Input: ImageInput{
+			Steps:           25,
+			Prompt:          request.Prompt,
+			Guidance:        3,
+			Seed:            int(time.Now().UnixNano()),
+			SafetyTolerance: 5,
+			NImages:         1, // replicate will always return 1 image
+			Width:           1440,
+			Height:          1440,
+			AspectRatio:     "1:1",
+		},
+	}, nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if !request.Stream {
+		// TODO: support non-stream mode
+		return nil, errors.Errorf("replicate models only support stream mode now, please set stream=true")
+	}
+
+	// Build the prompt from OpenAI messages
+	var promptBuilder strings.Builder
+	for _, message := range request.Messages {
+		switch msgCnt := message.Content.(type) {
+		case string:
+			promptBuilder.WriteString(message.Role)
+			promptBuilder.WriteString(": ")
+			promptBuilder.WriteString(msgCnt)
+			promptBuilder.WriteString("\n")
+		default:
+		}
+	}
+
+	replicateRequest := ReplicateChatRequest{
+		Input: ChatInput{
+			Prompt:           promptBuilder.String(),
+			MaxTokens:        request.MaxTokens,
+			Temperature:      1.0,
+			TopP:             1.0,
+			PresencePenalty:  0.0,
+			FrequencyPenalty: 0.0,
+		},
+	}
+
+	// Map optional fields
+	if request.Temperature != nil {
+		replicateRequest.Input.Temperature = *request.Temperature
+	}
+	if request.TopP != nil {
+		replicateRequest.Input.TopP = *request.TopP
+	}
+	if request.PresencePenalty != nil {
+		replicateRequest.Input.PresencePenalty = *request.PresencePenalty
+	}
+	if request.FrequencyPenalty != nil {
+		replicateRequest.Input.FrequencyPenalty = *request.FrequencyPenalty
+	}
+	if request.MaxTokens > 0 {
+		replicateRequest.Input.MaxTokens = request.MaxTokens
+	} else if request.MaxTokens == 0 {
+		replicateRequest.Input.MaxTokens = 500
+	}
+
+	return replicateRequest, nil
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	if !slices.Contains(ModelList, meta.OriginModelName) {
+		return "", errors.Errorf("model %s not supported", meta.OriginModelName)
+	}
+
+	return fmt.Sprintf("https://api.replicate.com/v1/models/%s/predictions", meta.OriginModelName), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
+	return nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	logger.Info(c, "send request to replicate")
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	switch meta.Mode {
+	case relaymode.ImagesGenerations:
+		err, usage = ImageHandler(c, resp)
+	case relaymode.ChatCompletions:
+		err, usage = ChatHandler(c, resp)
+	default:
+		err = openai.ErrorWrapper(errors.New("not implemented"), "not_implemented", http.StatusInternalServerError)
+	}
+
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "replicate"
+}
diff --git a/relay/adaptor/replicate/chat.go b/relay/adaptor/replicate/chat.go
new file mode 100644
index 00000000..4051f85c
--- /dev/null
+++ b/relay/adaptor/replicate/chat.go
@@ -0,0 +1,191 @@
+package replicate
+
+import (
+	"bufio"
+	"encoding/json"
+	"io"
+	"net/http"
+	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/render"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+func ChatHandler(c *gin.Context, resp *http.Response) (
+	srvErr *model.ErrorWithStatusCode, usage *model.Usage) {
+	if resp.StatusCode != http.StatusCreated {
+		payload, _ := io.ReadAll(resp.Body)
+		return openai.ErrorWrapper(
+				errors.Errorf("bad_status_code [%d]%s", resp.StatusCode, string(payload)),
+				"bad_status_code", http.StatusInternalServerError),
+			nil
+	}
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	respData := new(ChatResponse)
+	if err = json.Unmarshal(respBody, respData); err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	for {
+		err = func() error {
+			// get task
+			taskReq, err := http.NewRequestWithContext(c.Request.Context(),
+				http.MethodGet, respData.URLs.Get, nil)
+			if err != nil {
+				return errors.Wrap(err, "new request")
+			}
+
+			taskReq.Header.Set("Authorization", "Bearer "+meta.GetByContext(c).APIKey)
+			taskResp, err := http.DefaultClient.Do(taskReq)
+			if err != nil {
+				return errors.Wrap(err, "get task")
+			}
+			defer taskResp.Body.Close()
+
+			if taskResp.StatusCode != http.StatusOK {
+				payload, _ := io.ReadAll(taskResp.Body)
+				return errors.Errorf("bad status code [%d]%s",
+					taskResp.StatusCode, string(payload))
+			}
+
+			taskBody, err := io.ReadAll(taskResp.Body)
+			if err != nil {
+				return errors.Wrap(err, "read task response")
+			}
+
+			taskData := new(ChatResponse)
+			if err = json.Unmarshal(taskBody, taskData); err != nil {
+				return errors.Wrap(err, "decode task response")
+			}
+
+			switch taskData.Status {
+			case "succeeded":
+			case "failed", "canceled":
+				return errors.Errorf("task failed, [%s]%s", taskData.Status, taskData.Error)
+			default:
+				time.Sleep(time.Second * 3)
+				return errNextLoop
+			}
+
+			if taskData.URLs.Stream == "" {
+				return errors.New("stream url is empty")
+			}
+
+			// request stream url
+			responseText, err := chatStreamHandler(c, taskData.URLs.Stream)
+			if err != nil {
+				return errors.Wrap(err, "chat stream handler")
+			}
+
+			ctxMeta := meta.GetByContext(c)
+			usage = openai.ResponseText2Usage(responseText,
+				ctxMeta.ActualModelName, ctxMeta.PromptTokens)
+			return nil
+		}()
+		if err != nil {
+			if errors.Is(err, errNextLoop) {
+				continue
+			}
+
+			return openai.ErrorWrapper(err, "chat_task_failed", http.StatusInternalServerError), nil
+		}
+
+		break
+	}
+
+	return nil, usage
+}
+
+const (
+	eventPrefix = "event: "
+	dataPrefix  = "data: "
+	done        = "[DONE]"
+)
+
+func chatStreamHandler(c *gin.Context, streamUrl string) (responseText string, err error) {
+	// request stream endpoint
+	streamReq, err := http.NewRequestWithContext(c.Request.Context(), http.MethodGet, streamUrl, nil)
+	if err != nil {
+		return "", errors.Wrap(err, "new request to stream")
+	}
+
+	streamReq.Header.Set("Authorization", "Bearer "+meta.GetByContext(c).APIKey)
+	streamReq.Header.Set("Accept", "text/event-stream")
+	streamReq.Header.Set("Cache-Control", "no-store")
+
+	resp, err := http.DefaultClient.Do(streamReq)
+	if err != nil {
+		return "", errors.Wrap(err, "do request to stream")
+	}
+	defer resp.Body.Close()
+
+	if resp.StatusCode != http.StatusOK {
+		payload, _ := io.ReadAll(resp.Body)
+		return "", errors.Errorf("bad status code [%d]%s", resp.StatusCode, string(payload))
+	}
+
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(bufio.ScanLines)
+
+	common.SetEventStreamHeaders(c)
+	doneRendered := false
+	for scanner.Scan() {
+		line := strings.TrimSpace(scanner.Text())
+		if line == "" {
+			continue
+		}
+
+		// Handle comments starting with ':'
+		if strings.HasPrefix(line, ":") {
+			continue
+		}
+
+		// Parse SSE fields
+		if strings.HasPrefix(line, eventPrefix) {
+			event := strings.TrimSpace(line[len(eventPrefix):])
+			var data string
+			// Read the following lines to get data and id
+			for scanner.Scan() {
+				nextLine := scanner.Text()
+				if nextLine == "" {
+					break
+				}
+				if strings.HasPrefix(nextLine, dataPrefix) {
+					data = nextLine[len(dataPrefix):]
+				} else if strings.HasPrefix(nextLine, "id:") {
+					// id = strings.TrimSpace(nextLine[len("id:"):])
+				}
+			}
+
+			if event == "output" {
+				render.StringData(c, data)
+				responseText += data
+			} else if event == "done" {
+				render.Done(c)
+				doneRendered = true
+				break
+			}
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		return "", errors.Wrap(err, "scan stream")
+	}
+
+	if !doneRendered {
+		render.Done(c)
+	}
+
+	return responseText, nil
+}
diff --git a/relay/adaptor/replicate/constant.go b/relay/adaptor/replicate/constant.go
new file mode 100644
index 00000000..989142c9
--- /dev/null
+++ b/relay/adaptor/replicate/constant.go
@@ -0,0 +1,58 @@
+package replicate
+
+// ModelList is a list of models that can be used with Replicate.
+//
+// https://replicate.com/pricing
+var ModelList = []string{
+	// -------------------------------------
+	// image model
+	// -------------------------------------
+	"black-forest-labs/flux-1.1-pro",
+	"black-forest-labs/flux-1.1-pro-ultra",
+	"black-forest-labs/flux-canny-dev",
+	"black-forest-labs/flux-canny-pro",
+	"black-forest-labs/flux-depth-dev",
+	"black-forest-labs/flux-depth-pro",
+	"black-forest-labs/flux-dev",
+	"black-forest-labs/flux-dev-lora",
+	"black-forest-labs/flux-fill-dev",
+	"black-forest-labs/flux-fill-pro",
+	"black-forest-labs/flux-pro",
+	"black-forest-labs/flux-redux-dev",
+	"black-forest-labs/flux-redux-schnell",
+	"black-forest-labs/flux-schnell",
+	"black-forest-labs/flux-schnell-lora",
+	"ideogram-ai/ideogram-v2",
+	"ideogram-ai/ideogram-v2-turbo",
+	"recraft-ai/recraft-v3",
+	"recraft-ai/recraft-v3-svg",
+	"stability-ai/stable-diffusion-3",
+	"stability-ai/stable-diffusion-3.5-large",
+	"stability-ai/stable-diffusion-3.5-large-turbo",
+	"stability-ai/stable-diffusion-3.5-medium",
+	// -------------------------------------
+	// language model
+	// -------------------------------------
+	"ibm-granite/granite-20b-code-instruct-8k",
+	"ibm-granite/granite-3.0-2b-instruct",
+	"ibm-granite/granite-3.0-8b-instruct",
+	"ibm-granite/granite-8b-code-instruct-128k",
+	"meta/llama-2-13b",
+	"meta/llama-2-13b-chat",
+	"meta/llama-2-70b",
+	"meta/llama-2-70b-chat",
+	"meta/llama-2-7b",
+	"meta/llama-2-7b-chat",
+	"meta/meta-llama-3.1-405b-instruct",
+	"meta/meta-llama-3-70b",
+	"meta/meta-llama-3-70b-instruct",
+	"meta/meta-llama-3-8b",
+	"meta/meta-llama-3-8b-instruct",
+	"mistralai/mistral-7b-instruct-v0.2",
+	"mistralai/mistral-7b-v0.1",
+	"mistralai/mixtral-8x7b-instruct-v0.1",
+	// -------------------------------------
+	// video model
+	// -------------------------------------
+	// "minimax/video-01",  // TODO: implement the adaptor
+}
diff --git a/relay/adaptor/replicate/image.go b/relay/adaptor/replicate/image.go
new file mode 100644
index 00000000..3687249a
--- /dev/null
+++ b/relay/adaptor/replicate/image.go
@@ -0,0 +1,222 @@
+package replicate
+
+import (
+	"bytes"
+	"encoding/base64"
+	"encoding/json"
+	"fmt"
+	"image"
+	"image/png"
+	"io"
+	"net/http"
+	"sync"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"golang.org/x/image/webp"
+	"golang.org/x/sync/errgroup"
+)
+
+// ImagesEditsHandler just copy response body to client
+//
+// https://replicate.com/black-forest-labs/flux-fill-pro
+// func ImagesEditsHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+// 	c.Writer.WriteHeader(resp.StatusCode)
+// 	for k, v := range resp.Header {
+// 		c.Writer.Header().Set(k, v[0])
+// 	}
+
+// 	if _, err := io.Copy(c.Writer, resp.Body); err != nil {
+// 		return ErrorWrapper(err, "copy_response_body_failed", http.StatusInternalServerError), nil
+// 	}
+// 	defer resp.Body.Close()
+
+// 	return nil, nil
+// }
+
+var errNextLoop = errors.New("next_loop")
+
+func ImageHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	if resp.StatusCode != http.StatusCreated {
+		payload, _ := io.ReadAll(resp.Body)
+		return openai.ErrorWrapper(
+				errors.Errorf("bad_status_code [%d]%s", resp.StatusCode, string(payload)),
+				"bad_status_code", http.StatusInternalServerError),
+			nil
+	}
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	respData := new(ImageResponse)
+	if err = json.Unmarshal(respBody, respData); err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	for {
+		err = func() error {
+			// get task
+			taskReq, err := http.NewRequestWithContext(c.Request.Context(),
+				http.MethodGet, respData.URLs.Get, nil)
+			if err != nil {
+				return errors.Wrap(err, "new request")
+			}
+
+			taskReq.Header.Set("Authorization", "Bearer "+meta.GetByContext(c).APIKey)
+			taskResp, err := http.DefaultClient.Do(taskReq)
+			if err != nil {
+				return errors.Wrap(err, "get task")
+			}
+			defer taskResp.Body.Close()
+
+			if taskResp.StatusCode != http.StatusOK {
+				payload, _ := io.ReadAll(taskResp.Body)
+				return errors.Errorf("bad status code [%d]%s",
+					taskResp.StatusCode, string(payload))
+			}
+
+			taskBody, err := io.ReadAll(taskResp.Body)
+			if err != nil {
+				return errors.Wrap(err, "read task response")
+			}
+
+			taskData := new(ImageResponse)
+			if err = json.Unmarshal(taskBody, taskData); err != nil {
+				return errors.Wrap(err, "decode task response")
+			}
+
+			switch taskData.Status {
+			case "succeeded":
+			case "failed", "canceled":
+				return errors.Errorf("task failed: %s", taskData.Status)
+			default:
+				time.Sleep(time.Second * 3)
+				return errNextLoop
+			}
+
+			output, err := taskData.GetOutput()
+			if err != nil {
+				return errors.Wrap(err, "get output")
+			}
+			if len(output) == 0 {
+				return errors.New("response output is empty")
+			}
+
+			var mu sync.Mutex
+			var pool errgroup.Group
+			respBody := &openai.ImageResponse{
+				Created: taskData.CompletedAt.Unix(),
+				Data:    []openai.ImageData{},
+			}
+
+			for _, imgOut := range output {
+				imgOut := imgOut
+				pool.Go(func() error {
+					// download image
+					downloadReq, err := http.NewRequestWithContext(c.Request.Context(),
+						http.MethodGet, imgOut, nil)
+					if err != nil {
+						return errors.Wrap(err, "new request")
+					}
+
+					imgResp, err := http.DefaultClient.Do(downloadReq)
+					if err != nil {
+						return errors.Wrap(err, "download image")
+					}
+					defer imgResp.Body.Close()
+
+					if imgResp.StatusCode != http.StatusOK {
+						payload, _ := io.ReadAll(imgResp.Body)
+						return errors.Errorf("bad status code [%d]%s",
+							imgResp.StatusCode, string(payload))
+					}
+
+					imgData, err := io.ReadAll(imgResp.Body)
+					if err != nil {
+						return errors.Wrap(err, "read image")
+					}
+
+					imgData, err = ConvertImageToPNG(imgData)
+					if err != nil {
+						return errors.Wrap(err, "convert image")
+					}
+
+					mu.Lock()
+					respBody.Data = append(respBody.Data, openai.ImageData{
+						B64Json: fmt.Sprintf("data:image/png;base64,%s",
+							base64.StdEncoding.EncodeToString(imgData)),
+					})
+					mu.Unlock()
+
+					return nil
+				})
+			}
+
+			if err := pool.Wait(); err != nil {
+				if len(respBody.Data) == 0 {
+					return errors.WithStack(err)
+				}
+
+				logger.Error(c, fmt.Sprintf("some images failed to download: %+v", err))
+			}
+
+			c.JSON(http.StatusOK, respBody)
+			return nil
+		}()
+		if err != nil {
+			if errors.Is(err, errNextLoop) {
+				continue
+			}
+
+			return openai.ErrorWrapper(err, "image_task_failed", http.StatusInternalServerError), nil
+		}
+
+		break
+	}
+
+	return nil, nil
+}
+
+// ConvertImageToPNG converts a WebP image to PNG format
+func ConvertImageToPNG(webpData []byte) ([]byte, error) {
+	// bypass if it's already a PNG image
+	if bytes.HasPrefix(webpData, []byte("\x89PNG")) {
+		return webpData, nil
+	}
+
+	// check if is jpeg, convert to png
+	if bytes.HasPrefix(webpData, []byte("\xff\xd8\xff")) {
+		img, _, err := image.Decode(bytes.NewReader(webpData))
+		if err != nil {
+			return nil, errors.Wrap(err, "decode jpeg")
+		}
+
+		var pngBuffer bytes.Buffer
+		if err := png.Encode(&pngBuffer, img); err != nil {
+			return nil, errors.Wrap(err, "encode png")
+		}
+
+		return pngBuffer.Bytes(), nil
+	}
+
+	// Decode the WebP image
+	img, err := webp.Decode(bytes.NewReader(webpData))
+	if err != nil {
+		return nil, errors.Wrap(err, "decode webp")
+	}
+
+	// Encode the image as PNG
+	var pngBuffer bytes.Buffer
+	if err := png.Encode(&pngBuffer, img); err != nil {
+		return nil, errors.Wrap(err, "encode png")
+	}
+
+	return pngBuffer.Bytes(), nil
+}
diff --git a/relay/adaptor/replicate/model.go b/relay/adaptor/replicate/model.go
new file mode 100644
index 00000000..dba277eb
--- /dev/null
+++ b/relay/adaptor/replicate/model.go
@@ -0,0 +1,159 @@
+package replicate
+
+import (
+	"time"
+
+	"github.com/pkg/errors"
+)
+
+// DrawImageRequest draw image by fluxpro
+//
+// https://replicate.com/black-forest-labs/flux-pro?prediction=kg1krwsdf9rg80ch1sgsrgq7h8&output=json
+type DrawImageRequest struct {
+	Input ImageInput `json:"input"`
+}
+
+// ImageInput is input of DrawImageByFluxProRequest
+//
+// https://replicate.com/black-forest-labs/flux-1.1-pro/api/schema
+type ImageInput struct {
+	Steps           int    `json:"steps" binding:"required,min=1"`
+	Prompt          string `json:"prompt" binding:"required,min=5"`
+	ImagePrompt     string `json:"image_prompt"`
+	Guidance        int    `json:"guidance" binding:"required,min=2,max=5"`
+	Interval        int    `json:"interval" binding:"required,min=1,max=4"`
+	AspectRatio     string `json:"aspect_ratio" binding:"required,oneof=1:1 16:9 2:3 3:2 4:5 5:4 9:16"`
+	SafetyTolerance int    `json:"safety_tolerance" binding:"required,min=1,max=5"`
+	Seed            int    `json:"seed"`
+	NImages         int    `json:"n_images" binding:"required,min=1,max=8"`
+	Width           int    `json:"width" binding:"required,min=256,max=1440"`
+	Height          int    `json:"height" binding:"required,min=256,max=1440"`
+}
+
+// InpaintingImageByFlusReplicateRequest is request to inpainting image by flux pro
+//
+// https://replicate.com/black-forest-labs/flux-fill-pro/api/schema
+type InpaintingImageByFlusReplicateRequest struct {
+	Input FluxInpaintingInput `json:"input"`
+}
+
+// FluxInpaintingInput is input of DrawImageByFluxProRequest
+//
+// https://replicate.com/black-forest-labs/flux-fill-pro/api/schema
+type FluxInpaintingInput struct {
+	Mask             string `json:"mask" binding:"required"`
+	Image            string `json:"image" binding:"required"`
+	Seed             int    `json:"seed"`
+	Steps            int    `json:"steps" binding:"required,min=1"`
+	Prompt           string `json:"prompt" binding:"required,min=5"`
+	Guidance         int    `json:"guidance" binding:"required,min=2,max=5"`
+	OutputFormat     string `json:"output_format"`
+	SafetyTolerance  int    `json:"safety_tolerance" binding:"required,min=1,max=5"`
+	PromptUnsampling bool   `json:"prompt_unsampling"`
+}
+
+// ImageResponse is response of DrawImageByFluxProRequest
+//
+// https://replicate.com/black-forest-labs/flux-pro?prediction=kg1krwsdf9rg80ch1sgsrgq7h8&output=json
+type ImageResponse struct {
+	CompletedAt time.Time        `json:"completed_at"`
+	CreatedAt   time.Time        `json:"created_at"`
+	DataRemoved bool             `json:"data_removed"`
+	Error       string           `json:"error"`
+	ID          string           `json:"id"`
+	Input       DrawImageRequest `json:"input"`
+	Logs        string           `json:"logs"`
+	Metrics     FluxMetrics      `json:"metrics"`
+	// Output could be `string` or `[]string`
+	Output    any       `json:"output"`
+	StartedAt time.Time `json:"started_at"`
+	Status    string    `json:"status"`
+	URLs      FluxURLs  `json:"urls"`
+	Version   string    `json:"version"`
+}
+
+func (r *ImageResponse) GetOutput() ([]string, error) {
+	switch v := r.Output.(type) {
+	case string:
+		return []string{v}, nil
+	case []string:
+		return v, nil
+	case nil:
+		return nil, nil
+	case []interface{}:
+		// convert []interface{} to []string
+		ret := make([]string, len(v))
+		for idx, vv := range v {
+			if vvv, ok := vv.(string); ok {
+				ret[idx] = vvv
+			} else {
+				return nil, errors.Errorf("unknown output type: [%T]%v", vv, vv)
+			}
+		}
+
+		return ret, nil
+	default:
+		return nil, errors.Errorf("unknown output type: [%T]%v", r.Output, r.Output)
+	}
+}
+
+// FluxMetrics is metrics of ImageResponse
+type FluxMetrics struct {
+	ImageCount  int     `json:"image_count"`
+	PredictTime float64 `json:"predict_time"`
+	TotalTime   float64 `json:"total_time"`
+}
+
+// FluxURLs is urls of ImageResponse
+type FluxURLs struct {
+	Get    string `json:"get"`
+	Cancel string `json:"cancel"`
+}
+
+type ReplicateChatRequest struct {
+	Input ChatInput `json:"input" form:"input" binding:"required"`
+}
+
+// ChatInput is input of ChatByReplicateRequest
+//
+// https://replicate.com/meta/meta-llama-3.1-405b-instruct/api/schema
+type ChatInput struct {
+	TopK             int     `json:"top_k"`
+	TopP             float64 `json:"top_p"`
+	Prompt           string  `json:"prompt"`
+	MaxTokens        int     `json:"max_tokens"`
+	MinTokens        int     `json:"min_tokens"`
+	Temperature      float64 `json:"temperature"`
+	SystemPrompt     string  `json:"system_prompt"`
+	StopSequences    string  `json:"stop_sequences"`
+	PromptTemplate   string  `json:"prompt_template"`
+	PresencePenalty  float64 `json:"presence_penalty"`
+	FrequencyPenalty float64 `json:"frequency_penalty"`
+}
+
+// ChatResponse is response of ChatByReplicateRequest
+//
+// https://replicate.com/meta/meta-llama-3.1-405b-instruct/examples?input=http&output=json
+type ChatResponse struct {
+	CompletedAt time.Time   `json:"completed_at"`
+	CreatedAt   time.Time   `json:"created_at"`
+	DataRemoved bool        `json:"data_removed"`
+	Error       string      `json:"error"`
+	ID          string      `json:"id"`
+	Input       ChatInput   `json:"input"`
+	Logs        string      `json:"logs"`
+	Metrics     FluxMetrics `json:"metrics"`
+	// Output could be `string` or `[]string`
+	Output    []string        `json:"output"`
+	StartedAt time.Time       `json:"started_at"`
+	Status    string          `json:"status"`
+	URLs      ChatResponseUrl `json:"urls"`
+	Version   string          `json:"version"`
+}
+
+// ChatResponseUrl is task urls of ChatResponse
+type ChatResponseUrl struct {
+	Stream string `json:"stream"`
+	Get    string `json:"get"`
+	Cancel string `json:"cancel"`
+}
diff --git a/relay/adaptor/siliconflow/constants.go b/relay/adaptor/siliconflow/constants.go
new file mode 100644
index 00000000..0bf54761
--- /dev/null
+++ b/relay/adaptor/siliconflow/constants.go
@@ -0,0 +1,36 @@
+package siliconflow
+
+// https://docs.siliconflow.cn/docs/getting-started
+
+var ModelList = []string{
+	"deepseek-ai/deepseek-llm-67b-chat",
+	"Qwen/Qwen1.5-14B-Chat",
+	"Qwen/Qwen1.5-7B-Chat",
+	"Qwen/Qwen1.5-110B-Chat",
+	"Qwen/Qwen1.5-32B-Chat",
+	"01-ai/Yi-1.5-6B-Chat",
+	"01-ai/Yi-1.5-9B-Chat-16K",
+	"01-ai/Yi-1.5-34B-Chat-16K",
+	"THUDM/chatglm3-6b",
+	"deepseek-ai/DeepSeek-V2-Chat",
+	"THUDM/glm-4-9b-chat",
+	"Qwen/Qwen2-72B-Instruct",
+	"Qwen/Qwen2-7B-Instruct",
+	"Qwen/Qwen2-57B-A14B-Instruct",
+	"deepseek-ai/DeepSeek-Coder-V2-Instruct",
+	"Qwen/Qwen2-1.5B-Instruct",
+	"internlm/internlm2_5-7b-chat",
+	"BAAI/bge-large-en-v1.5",
+	"BAAI/bge-large-zh-v1.5",
+	"Pro/Qwen/Qwen2-7B-Instruct",
+	"Pro/Qwen/Qwen2-1.5B-Instruct",
+	"Pro/Qwen/Qwen1.5-7B-Chat",
+	"Pro/THUDM/glm-4-9b-chat",
+	"Pro/THUDM/chatglm3-6b",
+	"Pro/01-ai/Yi-1.5-9B-Chat-16K",
+	"Pro/01-ai/Yi-1.5-6B-Chat",
+	"Pro/google/gemma-2-9b-it",
+	"Pro/internlm/internlm2_5-7b-chat",
+	"Pro/meta-llama/Meta-Llama-3-8B-Instruct",
+	"Pro/mistralai/Mistral-7B-Instruct-v0.2",
+}
diff --git a/relay/adaptor/stepfun/constants.go b/relay/adaptor/stepfun/constants.go
new file mode 100644
index 00000000..6a2346ca
--- /dev/null
+++ b/relay/adaptor/stepfun/constants.go
@@ -0,0 +1,13 @@
+package stepfun
+
+var ModelList = []string{
+	"step-1-8k",
+	"step-1-32k",
+	"step-1-128k",
+	"step-1-256k",
+	"step-1-flash",
+	"step-2-16k",
+	"step-1v-8k",
+	"step-1v-32k",
+	"step-1x-medium",
+}
diff --git a/relay/adaptor/tencent/adaptor.go b/relay/adaptor/tencent/adaptor.go
new file mode 100644
index 00000000..0de92d4a
--- /dev/null
+++ b/relay/adaptor/tencent/adaptor.go
@@ -0,0 +1,89 @@
+package tencent
+
+import (
+	"errors"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+	"strconv"
+	"strings"
+)
+
+// https://cloud.tencent.com/document/api/1729/101837
+
+type Adaptor struct {
+	Sign      string
+	Action    string
+	Version   string
+	Timestamp int64
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.Action = "ChatCompletions"
+	a.Version = "2023-09-01"
+	a.Timestamp = helper.GetTimestamp()
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	return meta.BaseURL + "/", nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	req.Header.Set("Authorization", a.Sign)
+	req.Header.Set("X-TC-Action", a.Action)
+	req.Header.Set("X-TC-Version", a.Version)
+	req.Header.Set("X-TC-Timestamp", strconv.FormatInt(a.Timestamp, 10))
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	apiKey := c.Request.Header.Get("Authorization")
+	apiKey = strings.TrimPrefix(apiKey, "Bearer ")
+	_, secretId, secretKey, err := ParseConfig(apiKey)
+	if err != nil {
+		return nil, err
+	}
+	tencentRequest := ConvertRequest(*request)
+	// we have to calculate the sign here
+	a.Sign = GetSign(*tencentRequest, a, secretId, secretKey)
+	return tencentRequest, nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText = StreamHandler(c, resp)
+		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		err, usage = Handler(c, resp)
+	}
+	return
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "tencent"
+}
diff --git a/relay/adaptor/tencent/constants.go b/relay/adaptor/tencent/constants.go
new file mode 100644
index 00000000..e8631e5f
--- /dev/null
+++ b/relay/adaptor/tencent/constants.go
@@ -0,0 +1,9 @@
+package tencent
+
+var ModelList = []string{
+	"hunyuan-lite",
+	"hunyuan-standard",
+	"hunyuan-standard-256K",
+	"hunyuan-pro",
+	"hunyuan-vision",
+}
diff --git a/relay/channel/tencent/main.go b/relay/adaptor/tencent/main.go
similarity index 53%
rename from relay/channel/tencent/main.go
rename to relay/adaptor/tencent/main.go
index cfdc0bfd..827c8a46 100644
--- a/relay/channel/tencent/main.go
+++ b/relay/adaptor/tencent/main.go
@@ -3,48 +3,44 @@ package tencent
 import (
 	"bufio"
 	"crypto/hmac"
-	"crypto/sha1"
-	"encoding/base64"
+	"crypto/sha256"
+	"encoding/hex"
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
-	"sort"
 	"strconv"
 	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/conv"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
-// https://cloud.tencent.com/document/product/1729/97732
-
 func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
-	messages := make([]Message, 0, len(request.Messages))
+	messages := make([]*Message, 0, len(request.Messages))
 	for i := 0; i < len(request.Messages); i++ {
 		message := request.Messages[i]
-		messages = append(messages, Message{
+		messages = append(messages, &Message{
 			Content: message.StringContent(),
 			Role:    message.Role,
 		})
 	}
-	stream := 0
-	if request.Stream {
-		stream = 1
-	}
 	return &ChatRequest{
-		Timestamp:   helper.GetTimestamp(),
-		Expired:     helper.GetTimestamp() + 24*60*60,
-		QueryID:     helper.GetUUID(),
-		Temperature: request.Temperature,
-		TopP:        request.TopP,
-		Stream:      stream,
+		Model:       &request.Model,
+		Stream:      &request.Stream,
 		Messages:    messages,
+		TopP:        request.TopP,
+		Temperature: request.Temperature,
 	}
 }
 
@@ -52,7 +48,11 @@ func responseTencent2OpenAI(response *ChatResponse) *openai.TextResponse {
 	fullTextResponse := openai.TextResponse{
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
-		Usage:   response.Usage,
+		Usage: model.Usage{
+			PromptTokens:     response.Usage.PromptTokens,
+			CompletionTokens: response.Usage.CompletionTokens,
+			TotalTokens:      response.Usage.TotalTokens,
+		},
 	}
 	if len(response.Choices) > 0 {
 		choice := openai.TextResponseChoice{
@@ -70,7 +70,7 @@ func responseTencent2OpenAI(response *ChatResponse) *openai.TextResponse {
 
 func streamResponseTencent2OpenAI(TencentResponse *ChatResponse) *openai.ChatCompletionsStreamResponse {
 	response := openai.ChatCompletionsStreamResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   "tencent-hunyuan",
@@ -89,69 +89,52 @@ func streamResponseTencent2OpenAI(TencentResponse *ChatResponse) *openai.ChatCom
 func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, string) {
 	var responseText string
 	scanner := bufio.NewScanner(resp.Body)
-	scanner.Split(func(data []byte, atEOF bool) (advance int, token []byte, err error) {
-		if atEOF && len(data) == 0 {
-			return 0, nil, nil
-		}
-		if i := strings.Index(string(data), "\n"); i >= 0 {
-			return i + 1, data[0:i], nil
-		}
-		if atEOF {
-			return len(data), data, nil
-		}
-		return 0, nil, nil
-	})
-	dataChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			if len(data) < 5 { // ignore blank line or wrong format
-				continue
-			}
-			if data[:5] != "data:" {
-				continue
-			}
-			data = data[5:]
-			dataChan <- data
-		}
-		stopChan <- true
-	}()
+	scanner.Split(bufio.ScanLines)
+
 	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			var TencentResponse ChatResponse
-			err := json.Unmarshal([]byte(data), &TencentResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			response := streamResponseTencent2OpenAI(&TencentResponse)
-			if len(response.Choices) != 0 {
-				responseText += response.Choices[0].Delta.Content
-			}
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		if len(data) < 5 || !strings.HasPrefix(data, "data:") {
+			continue
 		}
-	})
+		data = strings.TrimPrefix(data, "data:")
+
+		var tencentResponse ChatResponse
+		err := json.Unmarshal([]byte(data), &tencentResponse)
+		if err != nil {
+			logger.SysError("error unmarshalling stream response: " + err.Error())
+			continue
+		}
+
+		response := streamResponseTencent2OpenAI(&tencentResponse)
+		if len(response.Choices) != 0 {
+			responseText += conv.AsString(response.Choices[0].Delta.Content)
+		}
+
+		err = render.ObjectData(c, response)
+		if err != nil {
+			logger.SysError(err.Error())
+		}
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), ""
 	}
+
 	return nil, responseText
 }
 
 func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
 	var TencentResponse ChatResponse
+	var responseP ChatResponseP
 	responseBody, err := io.ReadAll(resp.Body)
 	if err != nil {
 		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
@@ -160,10 +143,11 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
-	err = json.Unmarshal(responseBody, &TencentResponse)
+	err = json.Unmarshal(responseBody, &responseP)
 	if err != nil {
 		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
 	}
+	TencentResponse = responseP.Response
 	if TencentResponse.Error.Code != 0 {
 		return &model.ErrorWithStatusCode{
 			Error: model.Error{
@@ -200,29 +184,62 @@ func ParseConfig(config string) (appId int64, secretId string, secretKey string,
 	return
 }
 
-func GetSign(req ChatRequest, secretKey string) string {
-	params := make([]string, 0)
-	params = append(params, "app_id="+strconv.FormatInt(req.AppId, 10))
-	params = append(params, "secret_id="+req.SecretId)
-	params = append(params, "timestamp="+strconv.FormatInt(req.Timestamp, 10))
-	params = append(params, "query_id="+req.QueryID)
-	params = append(params, "temperature="+strconv.FormatFloat(req.Temperature, 'f', -1, 64))
-	params = append(params, "top_p="+strconv.FormatFloat(req.TopP, 'f', -1, 64))
-	params = append(params, "stream="+strconv.Itoa(req.Stream))
-	params = append(params, "expired="+strconv.FormatInt(req.Expired, 10))
-
-	var messageStr string
-	for _, msg := range req.Messages {
-		messageStr += fmt.Sprintf(`{"role":"%s","content":"%s"},`, msg.Role, msg.Content)
-	}
-	messageStr = strings.TrimSuffix(messageStr, ",")
-	params = append(params, "messages=["+messageStr+"]")
-
-	sort.Strings(params)
-	url := "hunyuan.cloud.tencent.com/hyllm/v1/chat/completions?" + strings.Join(params, "&")
-	mac := hmac.New(sha1.New, []byte(secretKey))
-	signURL := url
-	mac.Write([]byte(signURL))
-	sign := mac.Sum([]byte(nil))
-	return base64.StdEncoding.EncodeToString(sign)
+func sha256hex(s string) string {
+	b := sha256.Sum256([]byte(s))
+	return hex.EncodeToString(b[:])
+}
+
+func hmacSha256(s, key string) string {
+	hashed := hmac.New(sha256.New, []byte(key))
+	hashed.Write([]byte(s))
+	return string(hashed.Sum(nil))
+}
+
+func GetSign(req ChatRequest, adaptor *Adaptor, secId, secKey string) string {
+	// build canonical request string
+	host := "hunyuan.tencentcloudapi.com"
+	httpRequestMethod := "POST"
+	canonicalURI := "/"
+	canonicalQueryString := ""
+	canonicalHeaders := fmt.Sprintf("content-type:%s\nhost:%s\nx-tc-action:%s\n",
+		"application/json", host, strings.ToLower(adaptor.Action))
+	signedHeaders := "content-type;host;x-tc-action"
+	payload, _ := json.Marshal(req)
+	hashedRequestPayload := sha256hex(string(payload))
+	canonicalRequest := fmt.Sprintf("%s\n%s\n%s\n%s\n%s\n%s",
+		httpRequestMethod,
+		canonicalURI,
+		canonicalQueryString,
+		canonicalHeaders,
+		signedHeaders,
+		hashedRequestPayload)
+	// build string to sign
+	algorithm := "TC3-HMAC-SHA256"
+	requestTimestamp := strconv.FormatInt(adaptor.Timestamp, 10)
+	timestamp, _ := strconv.ParseInt(requestTimestamp, 10, 64)
+	t := time.Unix(timestamp, 0).UTC()
+	// must be the format 2006-01-02, ref to package time for more info
+	date := t.Format("2006-01-02")
+	credentialScope := fmt.Sprintf("%s/%s/tc3_request", date, "hunyuan")
+	hashedCanonicalRequest := sha256hex(canonicalRequest)
+	string2sign := fmt.Sprintf("%s\n%s\n%s\n%s",
+		algorithm,
+		requestTimestamp,
+		credentialScope,
+		hashedCanonicalRequest)
+
+	// sign string
+	secretDate := hmacSha256(date, "TC3"+secKey)
+	secretService := hmacSha256("hunyuan", secretDate)
+	secretKey := hmacSha256("tc3_request", secretService)
+	signature := hex.EncodeToString([]byte(hmacSha256(string2sign, secretKey)))
+
+	// build authorization
+	authorization := fmt.Sprintf("%s Credential=%s/%s, SignedHeaders=%s, Signature=%s",
+		algorithm,
+		secId,
+		credentialScope,
+		signedHeaders,
+		signature)
+	return authorization
 }
diff --git a/relay/adaptor/tencent/model.go b/relay/adaptor/tencent/model.go
new file mode 100644
index 00000000..fb97724e
--- /dev/null
+++ b/relay/adaptor/tencent/model.go
@@ -0,0 +1,75 @@
+package tencent
+
+type Message struct {
+	Role    string `json:"Role"`
+	Content string `json:"Content"`
+}
+
+type ChatRequest struct {
+	// 模型名称，可选值包括 hunyuan-lite、hunyuan-standard、hunyuan-standard-256K、hunyuan-pro。
+	// 各模型介绍请阅读 [产品概述](https://cloud.tencent.com/document/product/1729/104753) 中的说明。
+	//
+	// 注意：
+	// 不同的模型计费不同，请根据 [购买指南](https://cloud.tencent.com/document/product/1729/97731) 按需调用。
+	Model *string `json:"Model"`
+	// 聊天上下文信息。
+	// 说明：
+	// 1. 长度最多为 40，按对话时间从旧到新在数组中排列。
+	// 2. Message.Role 可选值：system、user、assistant。
+	// 其中，system 角色可选，如存在则必须位于列表的最开始。user 和 assistant 需交替出现（一问一答），以 user 提问开始和结束，且 Content 不能为空。Role 的顺序示例：[system（可选） user assistant user assistant user ...]。
+	// 3. Messages 中 Content 总长度不能超过模型输入长度上限（可参考 [产品概述](https://cloud.tencent.com/document/product/1729/104753) 文档），超过则会截断最前面的内容，只保留尾部内容。
+	Messages []*Message `json:"Messages"`
+	// 流式调用开关。
+	// 说明：
+	// 1. 未传值时默认为非流式调用（false）。
+	// 2. 流式调用时以 SSE 协议增量返回结果（返回值取 Choices[n].Delta 中的值，需要拼接增量数据才能获得完整结果）。
+	// 3. 非流式调用时：
+	// 调用方式与普通 HTTP 请求无异。
+	// 接口响应耗时较长，**如需更低时延建议设置为 true**。
+	// 只返回一次最终结果（返回值取 Choices[n].Message 中的值）。
+	//
+	// 注意：
+	// 通过 SDK 调用时，流式和非流式调用需用**不同的方式**获取返回值，具体参考 SDK 中的注释或示例（在各语言 SDK 代码仓库的 examples/hunyuan/v20230901/ 目录中）。
+	Stream *bool `json:"Stream"`
+	// 说明：
+	// 1. 影响输出文本的多样性，取值越大，生成文本的多样性越强。
+	// 2. 取值区间为 [0.0, 1.0]，未传值时使用各模型推荐值。
+	// 3. 非必要不建议使用，不合理的取值会影响效果。
+	TopP *float64 `json:"TopP"`
+	// 说明：
+	// 1. 较高的数值会使输出更加随机，而较低的数值会使其更加集中和确定。
+	// 2. 取值区间为 [0.0, 2.0]，未传值时使用各模型推荐值。
+	// 3. 非必要不建议使用，不合理的取值会影响效果。
+	Temperature *float64 `json:"Temperature"`
+}
+
+type Error struct {
+	Code    int    `json:"Code"`
+	Message string `json:"Message"`
+}
+
+type Usage struct {
+	PromptTokens     int `json:"PromptTokens"`
+	CompletionTokens int `json:"CompletionTokens"`
+	TotalTokens      int `json:"TotalTokens"`
+}
+
+type ResponseChoices struct {
+	FinishReason string  `json:"FinishReason,omitempty"` // 流式结束标志位，为 stop 则表示尾包
+	Messages     Message `json:"Message,omitempty"`      // 内容，同步模式返回内容，流模式为 null 输出 content 内容总数最多支持 1024token。
+	Delta        Message `json:"Delta,omitempty"`        // 内容，流模式返回内容，同步模式为 null 输出 content 内容总数最多支持 1024token。
+}
+
+type ChatResponse struct {
+	Choices []ResponseChoices `json:"Choices,omitempty"` // 结果
+	Created int64             `json:"Created,omitempty"` // unix 时间戳的字符串
+	Id      string            `json:"Id,omitempty"`      // 会话 id
+	Usage   Usage             `json:"Usage,omitempty"`   // token 数量
+	Error   Error             `json:"Error,omitempty"`   // 错误信息 注意：此字段可能返回 null，表示取不到有效值
+	Note    string            `json:"Note,omitempty"`    // 注释
+	ReqID   string            `json:"Req_id,omitempty"`  // 唯一请求 Id，每次请求都会返回。用于反馈接口入参
+}
+
+type ChatResponseP struct {
+	Response ChatResponse `json:"Response,omitempty"`
+}
diff --git a/relay/adaptor/togetherai/constants.go b/relay/adaptor/togetherai/constants.go
new file mode 100644
index 00000000..0a79fbdc
--- /dev/null
+++ b/relay/adaptor/togetherai/constants.go
@@ -0,0 +1,10 @@
+package togetherai
+
+// https://docs.together.ai/docs/inference-models
+
+var ModelList = []string{
+	"meta-llama/Llama-3-70b-chat-hf",
+	"deepseek-ai/deepseek-coder-33b-instruct",
+	"mistralai/Mixtral-8x22B-Instruct-v0.1",
+	"Qwen/Qwen1.5-72B-Chat",
+}
diff --git a/relay/adaptor/vertexai/adaptor.go b/relay/adaptor/vertexai/adaptor.go
new file mode 100644
index 00000000..3fab4a45
--- /dev/null
+++ b/relay/adaptor/vertexai/adaptor.go
@@ -0,0 +1,117 @@
+package vertexai
+
+import (
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	channelhelper "github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+var _ adaptor.Adaptor = new(Adaptor)
+
+const channelName = "vertexai"
+
+type Adaptor struct{}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	adaptor := GetAdaptor(request.Model)
+	if adaptor == nil {
+		return nil, errors.New("adaptor not found")
+	}
+
+	return adaptor.ConvertRequest(c, relayMode, request)
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	adaptor := GetAdaptor(meta.ActualModelName)
+	if adaptor == nil {
+		return nil, &relaymodel.ErrorWithStatusCode{
+			StatusCode: http.StatusInternalServerError,
+			Error: relaymodel.Error{
+				Message: "adaptor not found",
+			},
+		}
+	}
+	return adaptor.DoResponse(c, resp, meta)
+}
+
+func (a *Adaptor) GetModelList() (models []string) {
+	models = modelList
+	return
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return channelName
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	suffix := ""
+	if strings.HasPrefix(meta.ActualModelName, "gemini") {
+		if meta.IsStream {
+			suffix = "streamGenerateContent?alt=sse"
+		} else {
+			suffix = "generateContent"
+		}
+	} else {
+		if meta.IsStream {
+			suffix = "streamRawPredict?alt=sse"
+		} else {
+			suffix = "rawPredict"
+		}
+	}
+
+	if meta.BaseURL != "" {
+		return fmt.Sprintf(
+			"%s/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
+			meta.BaseURL,
+			meta.Config.VertexAIProjectID,
+			meta.Config.Region,
+			meta.ActualModelName,
+			suffix,
+		), nil
+	}
+	return fmt.Sprintf(
+		"https://%s-aiplatform.googleapis.com/v1/projects/%s/locations/%s/publishers/google/models/%s:%s",
+		meta.Config.Region,
+		meta.Config.VertexAIProjectID,
+		meta.Config.Region,
+		meta.ActualModelName,
+		suffix,
+	), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	token, err := getToken(c, meta.ChannelId, meta.Config.VertexAIADC)
+	if err != nil {
+		return err
+	}
+	req.Header.Set("Authorization", "Bearer "+token)
+	return nil
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
+}
diff --git a/relay/adaptor/vertexai/claude/adapter.go b/relay/adaptor/vertexai/claude/adapter.go
new file mode 100644
index 00000000..cb911cfe
--- /dev/null
+++ b/relay/adaptor/vertexai/claude/adapter.go
@@ -0,0 +1,60 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var ModelList = []string{
+	"claude-3-haiku@20240307",
+	"claude-3-sonnet@20240229",
+	"claude-3-opus@20240229",
+	"claude-3-5-sonnet@20240620",
+	"claude-3-5-sonnet-v2@20241022",
+	"claude-3-5-haiku@20241022",
+}
+
+const anthropicVersion = "vertex-2023-10-16"
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	claudeReq := anthropic.ConvertRequest(*request)
+	req := Request{
+		AnthropicVersion: anthropicVersion,
+		// Model:            claudeReq.Model,
+		Messages:    claudeReq.Messages,
+		System:      claudeReq.System,
+		MaxTokens:   claudeReq.MaxTokens,
+		Temperature: claudeReq.Temperature,
+		TopP:        claudeReq.TopP,
+		TopK:        claudeReq.TopK,
+		Stream:      claudeReq.Stream,
+		Tools:       claudeReq.Tools,
+	}
+
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, req)
+	return req, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, usage = anthropic.StreamHandler(c, resp)
+	} else {
+		err, usage = anthropic.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
diff --git a/relay/adaptor/vertexai/claude/model.go b/relay/adaptor/vertexai/claude/model.go
new file mode 100644
index 00000000..c08ba460
--- /dev/null
+++ b/relay/adaptor/vertexai/claude/model.go
@@ -0,0 +1,19 @@
+package vertexai
+
+import "github.com/songquanpeng/one-api/relay/adaptor/anthropic"
+
+type Request struct {
+	// AnthropicVersion must be "vertex-2023-10-16"
+	AnthropicVersion string `json:"anthropic_version"`
+	// Model            string              `json:"model"`
+	Messages      []anthropic.Message `json:"messages"`
+	System        string              `json:"system,omitempty"`
+	MaxTokens     int                 `json:"max_tokens,omitempty"`
+	StopSequences []string            `json:"stop_sequences,omitempty"`
+	Stream        bool                `json:"stream,omitempty"`
+	Temperature   *float64            `json:"temperature,omitempty"`
+	TopP          *float64            `json:"top_p,omitempty"`
+	TopK          int                 `json:"top_k,omitempty"`
+	Tools         []anthropic.Tool    `json:"tools,omitempty"`
+	ToolChoice    any                 `json:"tool_choice,omitempty"`
+}
diff --git a/relay/adaptor/vertexai/gemini/adapter.go b/relay/adaptor/vertexai/gemini/adapter.go
new file mode 100644
index 00000000..b5377875
--- /dev/null
+++ b/relay/adaptor/vertexai/gemini/adapter.go
@@ -0,0 +1,52 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/pkg/errors"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/relay/adaptor/gemini"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+var ModelList = []string{
+	"gemini-pro", "gemini-pro-vision",
+	"gemini-1.5-pro-001", "gemini-1.5-flash-001",
+	"gemini-1.5-pro-002", "gemini-1.5-flash-002",
+	"gemini-2.0-flash-exp", "gemini-2.0-flash-thinking-exp",
+}
+
+type Adaptor struct {
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+
+	geminiRequest := gemini.ConvertRequest(*request)
+	c.Set(ctxkey.RequestModel, request.Model)
+	c.Set(ctxkey.ConvertedRequest, geminiRequest)
+	return geminiRequest, nil
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		var responseText string
+		err, responseText = gemini.StreamHandler(c, resp)
+		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
+	} else {
+		switch meta.Mode {
+		case relaymode.Embeddings:
+			err, usage = gemini.EmbeddingHandler(c, resp)
+		default:
+			err, usage = gemini.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+		}
+	}
+	return
+}
diff --git a/relay/adaptor/vertexai/registry.go b/relay/adaptor/vertexai/registry.go
new file mode 100644
index 00000000..41099f02
--- /dev/null
+++ b/relay/adaptor/vertexai/registry.go
@@ -0,0 +1,50 @@
+package vertexai
+
+import (
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	claude "github.com/songquanpeng/one-api/relay/adaptor/vertexai/claude"
+	gemini "github.com/songquanpeng/one-api/relay/adaptor/vertexai/gemini"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+)
+
+type VertexAIModelType int
+
+const (
+	VerterAIClaude VertexAIModelType = iota + 1
+	VerterAIGemini
+)
+
+var modelMapping = map[string]VertexAIModelType{}
+var modelList = []string{}
+
+func init() {
+	modelList = append(modelList, claude.ModelList...)
+	for _, model := range claude.ModelList {
+		modelMapping[model] = VerterAIClaude
+	}
+
+	modelList = append(modelList, gemini.ModelList...)
+	for _, model := range gemini.ModelList {
+		modelMapping[model] = VerterAIGemini
+	}
+}
+
+type innerAIAdapter interface {
+	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
+	DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode)
+}
+
+func GetAdaptor(model string) innerAIAdapter {
+	adaptorType := modelMapping[model]
+	switch adaptorType {
+	case VerterAIClaude:
+		return &claude.Adaptor{}
+	case VerterAIGemini:
+		return &gemini.Adaptor{}
+	default:
+		return nil
+	}
+}
diff --git a/relay/adaptor/vertexai/token.go b/relay/adaptor/vertexai/token.go
new file mode 100644
index 00000000..0a5e0aad
--- /dev/null
+++ b/relay/adaptor/vertexai/token.go
@@ -0,0 +1,62 @@
+package vertexai
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"time"
+
+	credentials "cloud.google.com/go/iam/credentials/apiv1"
+	"cloud.google.com/go/iam/credentials/apiv1/credentialspb"
+	"github.com/patrickmn/go-cache"
+	"google.golang.org/api/option"
+)
+
+type ApplicationDefaultCredentials struct {
+	Type                    string `json:"type"`
+	ProjectID               string `json:"project_id"`
+	PrivateKeyID            string `json:"private_key_id"`
+	PrivateKey              string `json:"private_key"`
+	ClientEmail             string `json:"client_email"`
+	ClientID                string `json:"client_id"`
+	AuthURI                 string `json:"auth_uri"`
+	TokenURI                string `json:"token_uri"`
+	AuthProviderX509CertURL string `json:"auth_provider_x509_cert_url"`
+	ClientX509CertURL       string `json:"client_x509_cert_url"`
+	UniverseDomain          string `json:"universe_domain"`
+}
+
+var Cache = cache.New(50*time.Minute, 55*time.Minute)
+
+const defaultScope = "https://www.googleapis.com/auth/cloud-platform"
+
+func getToken(ctx context.Context, channelId int, adcJson string) (string, error) {
+	cacheKey := fmt.Sprintf("vertexai-token-%d", channelId)
+	if token, found := Cache.Get(cacheKey); found {
+		return token.(string), nil
+	}
+	adc := &ApplicationDefaultCredentials{}
+	if err := json.Unmarshal([]byte(adcJson), adc); err != nil {
+		return "", fmt.Errorf("Failed to decode credentials file: %w", err)
+	}
+
+	c, err := credentials.NewIamCredentialsClient(ctx, option.WithCredentialsJSON([]byte(adcJson)))
+	if err != nil {
+		return "", fmt.Errorf("Failed to create client: %w", err)
+	}
+	defer c.Close()
+
+	req := &credentialspb.GenerateAccessTokenRequest{
+		// See https://pkg.go.dev/cloud.google.com/go/iam/credentials/apiv1/credentialspb#GenerateAccessTokenRequest.
+		Name:  fmt.Sprintf("projects/-/serviceAccounts/%s", adc.ClientEmail),
+		Scope: []string{defaultScope},
+	}
+	resp, err := c.GenerateAccessToken(ctx, req)
+	if err != nil {
+		return "", fmt.Errorf("Failed to generate access token: %w", err)
+	}
+	_ = resp
+
+	Cache.Set(cacheKey, resp.AccessToken, cache.DefaultExpiration)
+	return resp.AccessToken, nil
+}
diff --git a/relay/adaptor/xai/constants.go b/relay/adaptor/xai/constants.go
new file mode 100644
index 00000000..9082b999
--- /dev/null
+++ b/relay/adaptor/xai/constants.go
@@ -0,0 +1,5 @@
+package xai
+
+var ModelList = []string{
+	"grok-beta",
+}
diff --git a/relay/channel/xunfei/adaptor.go b/relay/adaptor/xunfei/adaptor.go
similarity index 54%
rename from relay/channel/xunfei/adaptor.go
rename to relay/adaptor/xunfei/adaptor.go
index 92d9d7d6..b5967f26 100644
--- a/relay/channel/xunfei/adaptor.go
+++ b/relay/adaptor/xunfei/adaptor.go
@@ -3,10 +3,10 @@ package xunfei
 import (
 	"errors"
 	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
 	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 	"strings"
@@ -14,18 +14,19 @@ import (
 
 type Adaptor struct {
 	request *model.GeneralOpenAIRequest
+	meta    *meta.Meta
 }
 
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
+func (a *Adaptor) Init(meta *meta.Meta) {
+	a.meta = meta
 }
 
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	return "", nil
 }
 
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
 	// check DoResponse for auth part
 	return nil
 }
@@ -38,14 +39,21 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.G
 	return nil, nil
 }
 
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return request, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
 	// xunfei's request is not http request, so we don't need to do anything here
 	dummyResp := &http.Response{}
 	dummyResp.StatusCode = http.StatusOK
 	return dummyResp, nil
 }
 
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
 	splits := strings.Split(meta.APIKey, "|")
 	if len(splits) != 3 {
 		return nil, openai.ErrorWrapper(errors.New("invalid auth"), "invalid_auth", http.StatusBadRequest)
@@ -53,10 +61,18 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.Rel
 	if a.request == nil {
 		return nil, openai.ErrorWrapper(errors.New("request is nil"), "request_is_nil", http.StatusBadRequest)
 	}
+	version := parseAPIVersionByModelName(meta.ActualModelName)
+	if version == "" {
+		version = a.meta.Config.APIVersion
+	}
+	if version == "" {
+		version = "v1.1"
+	}
+	a.meta.Config.APIVersion = version
 	if meta.IsStream {
-		err, usage = StreamHandler(c, *a.request, splits[0], splits[1], splits[2])
+		err, usage = StreamHandler(c, meta, *a.request, splits[0], splits[1], splits[2])
 	} else {
-		err, usage = Handler(c, *a.request, splits[0], splits[1], splits[2])
+		err, usage = Handler(c, meta, *a.request, splits[0], splits[1], splits[2])
 	}
 	return
 }
diff --git a/relay/channel/xunfei/constants.go b/relay/adaptor/xunfei/constants.go
similarity index 67%
rename from relay/channel/xunfei/constants.go
rename to relay/adaptor/xunfei/constants.go
index 31dcec71..5b82ac29 100644
--- a/relay/channel/xunfei/constants.go
+++ b/relay/adaptor/xunfei/constants.go
@@ -5,5 +5,8 @@ var ModelList = []string{
 	"SparkDesk-v1.1",
 	"SparkDesk-v2.1",
 	"SparkDesk-v3.1",
+	"SparkDesk-v3.1-128K",
 	"SparkDesk-v3.5",
+	"SparkDesk-v3.5-32K",
+	"SparkDesk-v4.0",
 }
diff --git a/relay/channel/xunfei/main.go b/relay/adaptor/xunfei/main.go
similarity index 69%
rename from relay/channel/xunfei/main.go
rename to relay/adaptor/xunfei/main.go
index f89aea2b..3984ba5a 100644
--- a/relay/channel/xunfei/main.go
+++ b/relay/adaptor/xunfei/main.go
@@ -5,20 +5,24 @@ import (
 	"crypto/sha256"
 	"encoding/base64"
 	"encoding/json"
+	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/gorilla/websocket"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
 	"io"
 	"net/http"
 	"net/url"
 	"strings"
 	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/gorilla/websocket"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/common/random"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 // https://console.xfyun.cn/services/cbm
@@ -39,9 +43,38 @@ func requestOpenAI2Xunfei(request model.GeneralOpenAIRequest, xunfeiAppId string
 	xunfeiRequest.Parameter.Chat.TopK = request.N
 	xunfeiRequest.Parameter.Chat.MaxTokens = request.MaxTokens
 	xunfeiRequest.Payload.Message.Text = messages
+
+	if strings.HasPrefix(domain, "generalv3") || domain == "4.0Ultra" {
+		functions := make([]model.Function, len(request.Tools))
+		for i, tool := range request.Tools {
+			functions[i] = tool.Function
+		}
+		xunfeiRequest.Payload.Functions = &Functions{
+			Text: functions,
+		}
+	}
+
 	return &xunfeiRequest
 }
 
+func getToolCalls(response *ChatResponse) []model.Tool {
+	var toolCalls []model.Tool
+	if len(response.Payload.Choices.Text) == 0 {
+		return toolCalls
+	}
+	item := response.Payload.Choices.Text[0]
+	if item.FunctionCall == nil {
+		return toolCalls
+	}
+	toolCall := model.Tool{
+		Id:       fmt.Sprintf("call_%s", random.GetUUID()),
+		Type:     "function",
+		Function: *item.FunctionCall,
+	}
+	toolCalls = append(toolCalls, toolCall)
+	return toolCalls
+}
+
 func responseXunfei2OpenAI(response *ChatResponse) *openai.TextResponse {
 	if len(response.Payload.Choices.Text) == 0 {
 		response.Payload.Choices.Text = []ChatResponseTextItem{
@@ -53,13 +86,14 @@ func responseXunfei2OpenAI(response *ChatResponse) *openai.TextResponse {
 	choice := openai.TextResponseChoice{
 		Index: 0,
 		Message: model.Message{
-			Role:    "assistant",
-			Content: response.Payload.Choices.Text[0].Content,
+			Role:      "assistant",
+			Content:   response.Payload.Choices.Text[0].Content,
+			ToolCalls: getToolCalls(response),
 		},
 		FinishReason: constant.StopFinishReason,
 	}
 	fullTextResponse := openai.TextResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion",
 		Created: helper.GetTimestamp(),
 		Choices: []openai.TextResponseChoice{choice},
@@ -78,11 +112,12 @@ func streamResponseXunfei2OpenAI(xunfeiResponse *ChatResponse) *openai.ChatCompl
 	}
 	var choice openai.ChatCompletionsStreamResponseChoice
 	choice.Delta.Content = xunfeiResponse.Payload.Choices.Text[0].Content
+	choice.Delta.ToolCalls = getToolCalls(xunfeiResponse)
 	if xunfeiResponse.Payload.Choices.Status == 2 {
 		choice.FinishReason = &constant.StopFinishReason
 	}
 	response := openai.ChatCompletionsStreamResponse{
-		Id:      fmt.Sprintf("chatcmpl-%s", helper.GetUUID()),
+		Id:      fmt.Sprintf("chatcmpl-%s", random.GetUUID()),
 		Object:  "chat.completion.chunk",
 		Created: helper.GetTimestamp(),
 		Model:   "SparkDesk",
@@ -117,11 +152,11 @@ func buildXunfeiAuthUrl(hostUrl string, apiKey, apiSecret string) string {
 	return callUrl
 }
 
-func StreamHandler(c *gin.Context, textRequest model.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*model.ErrorWithStatusCode, *model.Usage) {
-	domain, authUrl := getXunfeiAuthUrl(c, apiKey, apiSecret, textRequest.Model)
+func StreamHandler(c *gin.Context, meta *meta.Meta, textRequest model.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*model.ErrorWithStatusCode, *model.Usage) {
+	domain, authUrl := getXunfeiAuthUrl(meta.Config.APIVersion, apiKey, apiSecret)
 	dataChan, stopChan, err := xunfeiMakeRequest(textRequest, domain, authUrl, appId)
 	if err != nil {
-		return openai.ErrorWrapper(err, "make xunfei request err", http.StatusInternalServerError), nil
+		return openai.ErrorWrapper(err, "xunfei_request_failed", http.StatusInternalServerError), nil
 	}
 	common.SetEventStreamHeaders(c)
 	var usage model.Usage
@@ -147,11 +182,11 @@ func StreamHandler(c *gin.Context, textRequest model.GeneralOpenAIRequest, appId
 	return nil, &usage
 }
 
-func Handler(c *gin.Context, textRequest model.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*model.ErrorWithStatusCode, *model.Usage) {
-	domain, authUrl := getXunfeiAuthUrl(c, apiKey, apiSecret, textRequest.Model)
+func Handler(c *gin.Context, meta *meta.Meta, textRequest model.GeneralOpenAIRequest, appId string, apiSecret string, apiKey string) (*model.ErrorWithStatusCode, *model.Usage) {
+	domain, authUrl := getXunfeiAuthUrl(meta.Config.APIVersion, apiKey, apiSecret)
 	dataChan, stopChan, err := xunfeiMakeRequest(textRequest, domain, authUrl, appId)
 	if err != nil {
-		return openai.ErrorWrapper(err, "make xunfei request err", http.StatusInternalServerError), nil
+		return openai.ErrorWrapper(err, "xunfei_request_failed", http.StatusInternalServerError), nil
 	}
 	var usage model.Usage
 	var content string
@@ -171,11 +206,7 @@ func Handler(c *gin.Context, textRequest model.GeneralOpenAIRequest, appId strin
 		}
 	}
 	if len(xunfeiResponse.Payload.Choices.Text) == 0 {
-		xunfeiResponse.Payload.Choices.Text = []ChatResponseTextItem{
-			{
-				Content: "",
-			},
-		}
+		return openai.ErrorWrapper(errors.New("xunfei empty response detected"), "xunfei_empty_response_detected", http.StatusInternalServerError), nil
 	}
 	xunfeiResponse.Payload.Choices.Text[0].Content = content
 
@@ -202,15 +233,21 @@ func xunfeiMakeRequest(textRequest model.GeneralOpenAIRequest, domain, authUrl,
 	if err != nil {
 		return nil, nil, err
 	}
+	_, msg, err := conn.ReadMessage()
+	if err != nil {
+		return nil, nil, err
+	}
 
 	dataChan := make(chan ChatResponse)
 	stopChan := make(chan bool)
 	go func() {
 		for {
-			_, msg, err := conn.ReadMessage()
-			if err != nil {
-				logger.SysError("error reading stream response: " + err.Error())
-				break
+			if msg == nil {
+				_, msg, err = conn.ReadMessage()
+				if err != nil {
+					logger.SysError("error reading stream response: " + err.Error())
+					break
+				}
 			}
 			var response ChatResponse
 			err = json.Unmarshal(msg, &response)
@@ -218,6 +255,7 @@ func xunfeiMakeRequest(textRequest model.GeneralOpenAIRequest, domain, authUrl,
 				logger.SysError("error unmarshalling stream response: " + err.Error())
 				break
 			}
+			msg = nil
 			dataChan <- response
 			if response.Payload.Choices.Status == 2 {
 				err := conn.Close()
@@ -233,45 +271,47 @@ func xunfeiMakeRequest(textRequest model.GeneralOpenAIRequest, domain, authUrl,
 	return dataChan, stopChan, nil
 }
 
-func getAPIVersion(c *gin.Context, modelName string) string {
-	query := c.Request.URL.Query()
-	apiVersion := query.Get("api-version")
-	if apiVersion != "" {
-		return apiVersion
+func parseAPIVersionByModelName(modelName string) string {
+	index := strings.IndexAny(modelName, "-")
+	if index != -1 {
+		return modelName[index+1:]
 	}
-	parts := strings.Split(modelName, "-")
-	if len(parts) == 2 {
-		apiVersion = parts[1]
-		return apiVersion
-
-	}
-	apiVersion = c.GetString(common.ConfigKeyAPIVersion)
-	if apiVersion != "" {
-		return apiVersion
-	}
-	apiVersion = "v1.1"
-	logger.SysLog("api_version not found, using default: " + apiVersion)
-	return apiVersion
+	return ""
 }
 
 // https://www.xfyun.cn/doc/spark/Web.html#_1-%E6%8E%A5%E5%8F%A3%E8%AF%B4%E6%98%8E
 func apiVersion2domain(apiVersion string) string {
 	switch apiVersion {
 	case "v1.1":
-		return "general"
+		return "lite"
 	case "v2.1":
 		return "generalv2"
 	case "v3.1":
 		return "generalv3"
+	case "v3.1-128K":
+		return "pro-128k"
 	case "v3.5":
 		return "generalv3.5"
+	case "v3.5-32K":
+		return "max-32k"
+	case "v4.0":
+		return "4.0Ultra"
 	}
 	return "general" + apiVersion
 }
 
-func getXunfeiAuthUrl(c *gin.Context, apiKey string, apiSecret string, modelName string) (string, string) {
-	apiVersion := getAPIVersion(c, modelName)
+func getXunfeiAuthUrl(apiVersion string, apiKey string, apiSecret string) (string, string) {
+	var authUrl string
 	domain := apiVersion2domain(apiVersion)
-	authUrl := buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)
+	switch apiVersion {
+	case "v3.1-128K":
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/pro-128k"), apiKey, apiSecret)
+		break
+	case "v3.5-32K":
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/chat/max-32k"), apiKey, apiSecret)
+		break
+	default:
+		authUrl = buildXunfeiAuthUrl(fmt.Sprintf("wss://spark-api.xf-yun.com/%s/chat", apiVersion), apiKey, apiSecret)
+	}
 	return domain, authUrl
 }
diff --git a/relay/channel/xunfei/model.go b/relay/adaptor/xunfei/model.go
similarity index 66%
rename from relay/channel/xunfei/model.go
rename to relay/adaptor/xunfei/model.go
index 1266739d..c9fb1bb8 100644
--- a/relay/channel/xunfei/model.go
+++ b/relay/adaptor/xunfei/model.go
@@ -9,30 +9,37 @@ type Message struct {
 	Content string `json:"content"`
 }
 
+type Functions struct {
+	Text []model.Function `json:"text,omitempty"`
+}
+
 type ChatRequest struct {
 	Header struct {
 		AppId string `json:"app_id"`
 	} `json:"header"`
 	Parameter struct {
 		Chat struct {
-			Domain      string  `json:"domain,omitempty"`
-			Temperature float64 `json:"temperature,omitempty"`
-			TopK        int     `json:"top_k,omitempty"`
-			MaxTokens   int     `json:"max_tokens,omitempty"`
-			Auditing    bool    `json:"auditing,omitempty"`
+			Domain      string   `json:"domain,omitempty"`
+			Temperature *float64 `json:"temperature,omitempty"`
+			TopK        int      `json:"top_k,omitempty"`
+			MaxTokens   int      `json:"max_tokens,omitempty"`
+			Auditing    bool     `json:"auditing,omitempty"`
 		} `json:"chat"`
 	} `json:"parameter"`
 	Payload struct {
 		Message struct {
 			Text []Message `json:"text"`
 		} `json:"message"`
+		Functions *Functions `json:"functions,omitempty"`
 	} `json:"payload"`
 }
 
 type ChatResponseTextItem struct {
-	Content string `json:"content"`
-	Role    string `json:"role"`
-	Index   int    `json:"index"`
+	Content      string          `json:"content"`
+	Role         string          `json:"role"`
+	Index        int             `json:"index"`
+	ContentType  string          `json:"content_type"`
+	FunctionCall *model.Function `json:"function_call"`
 }
 
 type ChatResponse struct {
diff --git a/relay/adaptor/zhipu/adaptor.go b/relay/adaptor/zhipu/adaptor.go
new file mode 100644
index 00000000..660bd379
--- /dev/null
+++ b/relay/adaptor/zhipu/adaptor.go
@@ -0,0 +1,149 @@
+package zhipu
+
+import (
+	"errors"
+	"fmt"
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+	"io"
+	"net/http"
+	"strings"
+)
+
+type Adaptor struct {
+	APIVersion string
+}
+
+func (a *Adaptor) Init(meta *meta.Meta) {
+
+}
+
+func (a *Adaptor) SetVersionByModeName(modelName string) {
+	if strings.HasPrefix(modelName, "glm-") {
+		a.APIVersion = "v4"
+	} else {
+		a.APIVersion = "v3"
+	}
+}
+
+func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
+	switch meta.Mode {
+	case relaymode.ImagesGenerations:
+		return fmt.Sprintf("%s/api/paas/v4/images/generations", meta.BaseURL), nil
+	case relaymode.Embeddings:
+		return fmt.Sprintf("%s/api/paas/v4/embeddings", meta.BaseURL), nil
+	}
+	a.SetVersionByModeName(meta.ActualModelName)
+	if a.APIVersion == "v4" {
+		return fmt.Sprintf("%s/api/paas/v4/chat/completions", meta.BaseURL), nil
+	}
+	method := "invoke"
+	if meta.IsStream {
+		method = "sse-invoke"
+	}
+	return fmt.Sprintf("%s/api/paas/v3/model-api/%s/%s", meta.BaseURL, meta.ActualModelName, method), nil
+}
+
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *meta.Meta) error {
+	adaptor.SetupCommonRequestHeader(c, req, meta)
+	token := GetToken(meta.APIKey)
+	req.Header.Set("Authorization", token)
+	return nil
+}
+
+func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	switch relayMode {
+	case relaymode.Embeddings:
+		baiduEmbeddingRequest, err := ConvertEmbeddingRequest(*request)
+		return baiduEmbeddingRequest, err
+	default:
+		// TopP [0.0, 1.0]
+		request.TopP = helper.Float64PtrMax(request.TopP, 1)
+		request.TopP = helper.Float64PtrMin(request.TopP, 0)
+
+		// Temperature [0.0, 1.0]
+		request.Temperature = helper.Float64PtrMax(request.Temperature, 1)
+		request.Temperature = helper.Float64PtrMin(request.Temperature, 0)
+		a.SetVersionByModeName(request.Model)
+		if a.APIVersion == "v4" {
+			return request, nil
+		}
+		return ConvertRequest(*request), nil
+	}
+}
+
+func (a *Adaptor) ConvertImageRequest(request *model.ImageRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	newRequest := ImageRequest{
+		Model:  request.Model,
+		Prompt: request.Prompt,
+		UserId: request.User,
+	}
+	return newRequest, nil
+}
+
+func (a *Adaptor) DoRequest(c *gin.Context, meta *meta.Meta, requestBody io.Reader) (*http.Response, error) {
+	return adaptor.DoRequestHelper(a, c, meta, requestBody)
+}
+
+func (a *Adaptor) DoResponseV4(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	if meta.IsStream {
+		err, _, usage = openai.StreamHandler(c, resp, meta.Mode)
+	} else {
+		err, usage = openai.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
+	}
+	return
+}
+
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *meta.Meta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
+	switch meta.Mode {
+	case relaymode.Embeddings:
+		err, usage = EmbeddingsHandler(c, resp)
+		return
+	case relaymode.ImagesGenerations:
+		err, usage = openai.ImageHandler(c, resp)
+		return
+	}
+	if a.APIVersion == "v4" {
+		return a.DoResponseV4(c, resp, meta)
+	}
+	if meta.IsStream {
+		err, usage = StreamHandler(c, resp)
+	} else {
+		if meta.Mode == relaymode.Embeddings {
+			err, usage = EmbeddingsHandler(c, resp)
+		} else {
+			err, usage = Handler(c, resp)
+		}
+	}
+	return
+}
+
+func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) (*EmbeddingRequest, error) {
+	inputs := request.ParseInput()
+	if len(inputs) != 1 {
+		return nil, errors.New("invalid input length, zhipu only support one input")
+	}
+	return &EmbeddingRequest{
+		Model: request.Model,
+		Input: inputs[0],
+	}, nil
+}
+
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+func (a *Adaptor) GetChannelName() string {
+	return "zhipu"
+}
diff --git a/relay/channel/zhipu/constants.go b/relay/adaptor/zhipu/constants.go
similarity index 62%
rename from relay/channel/zhipu/constants.go
rename to relay/adaptor/zhipu/constants.go
index 1655a59d..e1192123 100644
--- a/relay/channel/zhipu/constants.go
+++ b/relay/adaptor/zhipu/constants.go
@@ -2,5 +2,6 @@ package zhipu
 
 var ModelList = []string{
 	"chatglm_turbo", "chatglm_pro", "chatglm_std", "chatglm_lite",
-	"glm-4", "glm-4v", "glm-3-turbo",
+	"glm-4", "glm-4v", "glm-3-turbo", "embedding-2",
+	"cogview-3",
 }
diff --git a/relay/channel/zhipu/main.go b/relay/adaptor/zhipu/main.go
similarity index 66%
rename from relay/channel/zhipu/main.go
rename to relay/adaptor/zhipu/main.go
index a46fd537..ab3a5678 100644
--- a/relay/channel/zhipu/main.go
+++ b/relay/adaptor/zhipu/main.go
@@ -3,19 +3,21 @@ package zhipu
 import (
 	"bufio"
 	"encoding/json"
-	"github.com/gin-gonic/gin"
-	"github.com/golang-jwt/jwt"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/helper"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
 	"strings"
 	"sync"
 	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/golang-jwt/jwt"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/helper"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 // https://open.bigmodel.cn/doc/api#chatglm_std
@@ -155,66 +157,55 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 		}
 		return 0, nil, nil
 	})
-	dataChan := make(chan string)
-	metaChan := make(chan string)
-	stopChan := make(chan bool)
-	go func() {
-		for scanner.Scan() {
-			data := scanner.Text()
-			lines := strings.Split(data, "\n")
-			for i, line := range lines {
-				if len(line) < 5 {
+
+	common.SetEventStreamHeaders(c)
+
+	for scanner.Scan() {
+		data := scanner.Text()
+		lines := strings.Split(data, "\n")
+		for i, line := range lines {
+			if len(line) < 5 {
+				continue
+			}
+			if strings.HasPrefix(line, "data:") {
+				dataSegment := line[5:]
+				if i != len(lines)-1 {
+					dataSegment += "\n"
+				}
+				response := streamResponseZhipu2OpenAI(dataSegment)
+				err := render.ObjectData(c, response)
+				if err != nil {
+					logger.SysError("error marshalling stream response: " + err.Error())
+				}
+			} else if strings.HasPrefix(line, "meta:") {
+				metaSegment := line[5:]
+				var zhipuResponse StreamMetaResponse
+				err := json.Unmarshal([]byte(metaSegment), &zhipuResponse)
+				if err != nil {
+					logger.SysError("error unmarshalling stream response: " + err.Error())
 					continue
 				}
-				if line[:5] == "data:" {
-					dataChan <- line[5:]
-					if i != len(lines)-1 {
-						dataChan <- "\n"
-					}
-				} else if line[:5] == "meta:" {
-					metaChan <- line[5:]
+				response, zhipuUsage := streamMetaResponseZhipu2OpenAI(&zhipuResponse)
+				err = render.ObjectData(c, response)
+				if err != nil {
+					logger.SysError("error marshalling stream response: " + err.Error())
 				}
+				usage = zhipuUsage
 			}
 		}
-		stopChan <- true
-	}()
-	common.SetEventStreamHeaders(c)
-	c.Stream(func(w io.Writer) bool {
-		select {
-		case data := <-dataChan:
-			response := streamResponseZhipu2OpenAI(data)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case data := <-metaChan:
-			var zhipuResponse StreamMetaResponse
-			err := json.Unmarshal([]byte(data), &zhipuResponse)
-			if err != nil {
-				logger.SysError("error unmarshalling stream response: " + err.Error())
-				return true
-			}
-			response, zhipuUsage := streamMetaResponseZhipu2OpenAI(&zhipuResponse)
-			jsonResponse, err := json.Marshal(response)
-			if err != nil {
-				logger.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			usage = zhipuUsage
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonResponse)})
-			return true
-		case <-stopChan:
-			c.Render(-1, common.CustomEvent{Data: "data: [DONE]"})
-			return false
-		}
-	})
+	}
+
+	if err := scanner.Err(); err != nil {
+		logger.SysError("error reading stream: " + err.Error())
+	}
+
+	render.Done(c)
+
 	err := resp.Body.Close()
 	if err != nil {
 		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 	}
+
 	return nil, usage
 }
 
@@ -254,3 +245,50 @@ func Handler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *
 	_, err = c.Writer.Write(jsonResponse)
 	return nil, &fullTextResponse.Usage
 }
+
+func EmbeddingsHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusCode, *model.Usage) {
+	var zhipuResponse EmbeddingResponse
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = json.Unmarshal(responseBody, &zhipuResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	fullTextResponse := embeddingResponseZhipu2OpenAI(&zhipuResponse)
+	jsonResponse, err := json.Marshal(fullTextResponse)
+	if err != nil {
+		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	return nil, &fullTextResponse.Usage
+}
+
+func embeddingResponseZhipu2OpenAI(response *EmbeddingResponse) *openai.EmbeddingResponse {
+	openAIEmbeddingResponse := openai.EmbeddingResponse{
+		Object: "list",
+		Data:   make([]openai.EmbeddingResponseItem, 0, len(response.Embeddings)),
+		Model:  response.Model,
+		Usage: model.Usage{
+			PromptTokens:     response.PromptTokens,
+			CompletionTokens: response.CompletionTokens,
+			TotalTokens:      response.Usage.TotalTokens,
+		},
+	}
+
+	for _, item := range response.Embeddings {
+		openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
+			Object:    `embedding`,
+			Index:     item.Index,
+			Embedding: item.Embedding,
+		})
+	}
+	return &openAIEmbeddingResponse
+}
diff --git a/relay/channel/zhipu/model.go b/relay/adaptor/zhipu/model.go
similarity index 59%
rename from relay/channel/zhipu/model.go
rename to relay/adaptor/zhipu/model.go
index b63e1d6f..06e22dc1 100644
--- a/relay/channel/zhipu/model.go
+++ b/relay/adaptor/zhipu/model.go
@@ -12,8 +12,8 @@ type Message struct {
 
 type Request struct {
 	Prompt      []Message `json:"prompt"`
-	Temperature float64   `json:"temperature,omitempty"`
-	TopP        float64   `json:"top_p,omitempty"`
+	Temperature *float64  `json:"temperature,omitempty"`
+	TopP        *float64  `json:"top_p,omitempty"`
 	RequestId   string    `json:"request_id,omitempty"`
 	Incremental bool      `json:"incremental,omitempty"`
 }
@@ -44,3 +44,27 @@ type tokenData struct {
 	Token      string
 	ExpiryTime time.Time
 }
+
+type EmbeddingRequest struct {
+	Model string `json:"model"`
+	Input string `json:"input"`
+}
+
+type EmbeddingResponse struct {
+	Model       string          `json:"model"`
+	Object      string          `json:"object"`
+	Embeddings  []EmbeddingData `json:"data"`
+	model.Usage `json:"usage"`
+}
+
+type EmbeddingData struct {
+	Index     int       `json:"index"`
+	Object    string    `json:"object"`
+	Embedding []float64 `json:"embedding"`
+}
+
+type ImageRequest struct {
+	Model  string `json:"model"`
+	Prompt string `json:"prompt"`
+	UserId string `json:"user_id,omitempty"`
+}
diff --git a/relay/adaptor_test.go b/relay/adaptor_test.go
new file mode 100644
index 00000000..884c8e9f
--- /dev/null
+++ b/relay/adaptor_test.go
@@ -0,0 +1,16 @@
+package relay
+
+import (
+	. "github.com/smartystreets/goconvey/convey"
+	"github.com/songquanpeng/one-api/relay/apitype"
+	"testing"
+)
+
+func TestGetAdaptor(t *testing.T) {
+	Convey("get adaptor", t, func() {
+		for i := 0; i < apitype.Dummy; i++ {
+			a := GetAdaptor(i)
+			So(a, ShouldNotBeNil)
+		}
+	})
+}
diff --git a/relay/apitype/define.go b/relay/apitype/define.go
new file mode 100644
index 00000000..0c6a5ff1
--- /dev/null
+++ b/relay/apitype/define.go
@@ -0,0 +1,25 @@
+package apitype
+
+const (
+	OpenAI = iota
+	Anthropic
+	PaLM
+	Baidu
+	Zhipu
+	Ali
+	Xunfei
+	AIProxyLibrary
+	Tencent
+	Gemini
+	Ollama
+	AwsClaude
+	Coze
+	Cohere
+	Cloudflare
+	DeepL
+	VertexAI
+	Proxy
+	Replicate
+
+	Dummy // this one is only for count, do not add any channel after this
+)
diff --git a/relay/billing/billing.go b/relay/billing/billing.go
new file mode 100644
index 00000000..a99d37ee
--- /dev/null
+++ b/relay/billing/billing.go
@@ -0,0 +1,42 @@
+package billing
+
+import (
+	"context"
+	"fmt"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
+)
+
+func ReturnPreConsumedQuota(ctx context.Context, preConsumedQuota int64, tokenId int) {
+	if preConsumedQuota != 0 {
+		go func(ctx context.Context) {
+			// return pre-consumed quota
+			err := model.PostConsumeTokenQuota(tokenId, -preConsumedQuota)
+			if err != nil {
+				logger.Error(ctx, "error return pre-consumed quota: "+err.Error())
+			}
+		}(ctx)
+	}
+}
+
+func PostConsumeQuota(ctx context.Context, tokenId int, quotaDelta int64, totalQuota int64, userId int, channelId int, modelRatio float64, groupRatio float64, modelName string, tokenName string) {
+	// quotaDelta is remaining quota to be consumed
+	err := model.PostConsumeTokenQuota(tokenId, quotaDelta)
+	if err != nil {
+		logger.SysError("error consuming token remain quota: " + err.Error())
+	}
+	err = model.CacheUpdateUserQuota(ctx, userId)
+	if err != nil {
+		logger.SysError("error update user quota cache: " + err.Error())
+	}
+	// totalQuota is total quota consumed
+	if totalQuota != 0 {
+		logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
+		model.RecordConsumeLog(ctx, userId, channelId, int(totalQuota), 0, modelName, tokenName, totalQuota, logContent)
+		model.UpdateUserUsedQuotaAndRequestCount(userId, totalQuota)
+		model.UpdateChannelUsedQuota(channelId, totalQuota)
+	}
+	if totalQuota <= 0 {
+		logger.Error(ctx, fmt.Sprintf("totalQuota consumed is %d, something is wrong", totalQuota))
+	}
+}
diff --git a/common/group-ratio.go b/relay/billing/ratio/group.go
similarity index 97%
rename from common/group-ratio.go
rename to relay/billing/ratio/group.go
index 2de6e810..8e9c5b73 100644
--- a/common/group-ratio.go
+++ b/relay/billing/ratio/group.go
@@ -1,4 +1,4 @@
-package common
+package ratio
 
 import (
 	"encoding/json"
diff --git a/relay/billing/ratio/image.go b/relay/billing/ratio/image.go
new file mode 100644
index 00000000..c8c42a15
--- /dev/null
+++ b/relay/billing/ratio/image.go
@@ -0,0 +1,66 @@
+package ratio
+
+var ImageSizeRatios = map[string]map[string]float64{
+	"dall-e-2": {
+		"256x256":   1,
+		"512x512":   1.125,
+		"1024x1024": 1.25,
+	},
+	"dall-e-3": {
+		"1024x1024": 1,
+		"1024x1792": 2,
+		"1792x1024": 2,
+	},
+	"ali-stable-diffusion-xl": {
+		"512x1024":  1,
+		"1024x768":  1,
+		"1024x1024": 1,
+		"576x1024":  1,
+		"1024x576":  1,
+	},
+	"ali-stable-diffusion-v1.5": {
+		"512x1024":  1,
+		"1024x768":  1,
+		"1024x1024": 1,
+		"576x1024":  1,
+		"1024x576":  1,
+	},
+	"wanx-v1": {
+		"1024x1024": 1,
+		"720x1280":  1,
+		"1280x720":  1,
+	},
+	"step-1x-medium": {
+		"256x256":   1,
+		"512x512":   1,
+		"768x768":   1,
+		"1024x1024": 1,
+		"1280x800":  1,
+		"800x1280":  1,
+	},
+}
+
+var ImageGenerationAmounts = map[string][2]int{
+	"dall-e-2":                  {1, 10},
+	"dall-e-3":                  {1, 1}, // OpenAI allows n=1 currently.
+	"ali-stable-diffusion-xl":   {1, 4}, // Ali
+	"ali-stable-diffusion-v1.5": {1, 4}, // Ali
+	"wanx-v1":                   {1, 4}, // Ali
+	"cogview-3":                 {1, 1},
+	"step-1x-medium":            {1, 1},
+}
+
+var ImagePromptLengthLimitations = map[string]int{
+	"dall-e-2":                  1000,
+	"dall-e-3":                  4000,
+	"ali-stable-diffusion-xl":   4000,
+	"ali-stable-diffusion-v1.5": 4000,
+	"wanx-v1":                   4000,
+	"cogview-3":                 833,
+	"step-1x-medium":            4000,
+}
+
+var ImageOriginModelName = map[string]string{
+	"ali-stable-diffusion-xl":   "stable-diffusion-xl",
+	"ali-stable-diffusion-v1.5": "stable-diffusion-v1.5",
+}
diff --git a/relay/billing/ratio/model.go b/relay/billing/ratio/model.go
new file mode 100644
index 00000000..f83aa70c
--- /dev/null
+++ b/relay/billing/ratio/model.go
@@ -0,0 +1,538 @@
+package ratio
+
+import (
+	"encoding/json"
+	"fmt"
+	"strings"
+
+	"github.com/songquanpeng/one-api/common/logger"
+)
+
+const (
+	USD2RMB = 7
+	USD     = 500 // $0.002 = 1 -> $1 = 500
+	RMB     = USD / USD2RMB
+)
+
+// ModelRatio
+// https://platform.openai.com/docs/models/model-endpoint-compatibility
+// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/Blfmc9dlf
+// https://openai.com/pricing
+// 1 === $0.002 / 1K tokens
+// 1 === ￥0.014 / 1k tokens
+var ModelRatio = map[string]float64{
+	// https://openai.com/pricing
+	"gpt-4":                   15,
+	"gpt-4-0314":              15,
+	"gpt-4-0613":              15,
+	"gpt-4-32k":               30,
+	"gpt-4-32k-0314":          30,
+	"gpt-4-32k-0613":          30,
+	"gpt-4-1106-preview":      5,     // $0.01 / 1K tokens
+	"gpt-4-0125-preview":      5,     // $0.01 / 1K tokens
+	"gpt-4-turbo-preview":     5,     // $0.01 / 1K tokens
+	"gpt-4-turbo":             5,     // $0.01 / 1K tokens
+	"gpt-4-turbo-2024-04-09":  5,     // $0.01 / 1K tokens
+	"gpt-4o":                  2.5,   // $0.005 / 1K tokens
+	"chatgpt-4o-latest":       2.5,   // $0.005 / 1K tokens
+	"gpt-4o-2024-05-13":       2.5,   // $0.005 / 1K tokens
+	"gpt-4o-2024-08-06":       1.25,  // $0.0025 / 1K tokens
+	"gpt-4o-2024-11-20":       1.25,  // $0.0025 / 1K tokens
+	"gpt-4o-mini":             0.075, // $0.00015 / 1K tokens
+	"gpt-4o-mini-2024-07-18":  0.075, // $0.00015 / 1K tokens
+	"gpt-4-vision-preview":    5,     // $0.01 / 1K tokens
+	"gpt-3.5-turbo":           0.25,  // $0.0005 / 1K tokens
+	"gpt-3.5-turbo-0301":      0.75,
+	"gpt-3.5-turbo-0613":      0.75,
+	"gpt-3.5-turbo-16k":       1.5, // $0.003 / 1K tokens
+	"gpt-3.5-turbo-16k-0613":  1.5,
+	"gpt-3.5-turbo-instruct":  0.75, // $0.0015 / 1K tokens
+	"gpt-3.5-turbo-1106":      0.5,  // $0.001 / 1K tokens
+	"gpt-3.5-turbo-0125":      0.25, // $0.0005 / 1K tokens
+	"o1":                      7.5,  // $15.00 / 1M input tokens
+	"o1-2024-12-17":           7.5,
+	"o1-preview":              7.5, // $15.00 / 1M input tokens
+	"o1-preview-2024-09-12":   7.5,
+	"o1-mini":                 1.5, // $3.00 / 1M input tokens
+	"o1-mini-2024-09-12":      1.5,
+	"davinci-002":             1,   // $0.002 / 1K tokens
+	"babbage-002":             0.2, // $0.0004 / 1K tokens
+	"text-ada-001":            0.2,
+	"text-babbage-001":        0.25,
+	"text-curie-001":          1,
+	"text-davinci-002":        10,
+	"text-davinci-003":        10,
+	"text-davinci-edit-001":   10,
+	"code-davinci-edit-001":   10,
+	"whisper-1":               15,  // $0.006 / minute -> $0.006 / 150 words -> $0.006 / 200 tokens -> $0.03 / 1k tokens
+	"tts-1":                   7.5, // $0.015 / 1K characters
+	"tts-1-1106":              7.5,
+	"tts-1-hd":                15, // $0.030 / 1K characters
+	"tts-1-hd-1106":           15,
+	"davinci":                 10,
+	"curie":                   10,
+	"babbage":                 10,
+	"ada":                     10,
+	"text-embedding-ada-002":  0.05,
+	"text-embedding-3-small":  0.01,
+	"text-embedding-3-large":  0.065,
+	"text-search-ada-doc-001": 10,
+	"text-moderation-stable":  0.1,
+	"text-moderation-latest":  0.1,
+	"dall-e-2":                0.02 * USD, // $0.016 - $0.020 / image
+	"dall-e-3":                0.04 * USD, // $0.040 - $0.120 / image
+	// https://www.anthropic.com/api#pricing
+	"claude-instant-1.2":         0.8 / 1000 * USD,
+	"claude-2.0":                 8.0 / 1000 * USD,
+	"claude-2.1":                 8.0 / 1000 * USD,
+	"claude-3-haiku-20240307":    0.25 / 1000 * USD,
+	"claude-3-5-haiku-20241022":  1.0 / 1000 * USD,
+	"claude-3-sonnet-20240229":   3.0 / 1000 * USD,
+	"claude-3-5-sonnet-20240620": 3.0 / 1000 * USD,
+	"claude-3-5-sonnet-20241022": 3.0 / 1000 * USD,
+	"claude-3-opus-20240229":     15.0 / 1000 * USD,
+	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/hlrk4akp7
+	"ERNIE-4.0-8K":       0.120 * RMB,
+	"ERNIE-3.5-8K":       0.012 * RMB,
+	"ERNIE-3.5-8K-0205":  0.024 * RMB,
+	"ERNIE-3.5-8K-1222":  0.012 * RMB,
+	"ERNIE-Bot-8K":       0.024 * RMB,
+	"ERNIE-3.5-4K-0205":  0.012 * RMB,
+	"ERNIE-Speed-8K":     0.004 * RMB,
+	"ERNIE-Speed-128K":   0.004 * RMB,
+	"ERNIE-Lite-8K-0922": 0.008 * RMB,
+	"ERNIE-Lite-8K-0308": 0.003 * RMB,
+	"ERNIE-Tiny-8K":      0.001 * RMB,
+	"BLOOMZ-7B":          0.004 * RMB,
+	"Embedding-V1":       0.002 * RMB,
+	"bge-large-zh":       0.002 * RMB,
+	"bge-large-en":       0.002 * RMB,
+	"tao-8k":             0.002 * RMB,
+	// https://ai.google.dev/pricing
+	"gemini-pro":                    1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	"gemini-1.0-pro":                1,
+	"gemini-1.5-pro":                1,
+	"gemini-1.5-pro-001":            1,
+	"gemini-1.5-flash":              1,
+	"gemini-1.5-flash-001":          1,
+	"gemini-2.0-flash-exp":          1,
+	"gemini-2.0-flash-thinking-exp": 1,
+	"aqa":                           1,
+	// https://open.bigmodel.cn/pricing
+	"glm-4":         0.1 * RMB,
+	"glm-4v":        0.1 * RMB,
+	"glm-3-turbo":   0.005 * RMB,
+	"embedding-2":   0.0005 * RMB,
+	"chatglm_turbo": 0.3572, // ￥0.005 / 1k tokens
+	"chatglm_pro":   0.7143, // ￥0.01 / 1k tokens
+	"chatglm_std":   0.3572, // ￥0.005 / 1k tokens
+	"chatglm_lite":  0.1429, // ￥0.002 / 1k tokens
+	"cogview-3":     0.25 * RMB,
+	// https://help.aliyun.com/zh/dashscope/developer-reference/tongyi-thousand-questions-metering-and-billing
+	"qwen-turbo":                  1.4286, // ￥0.02 / 1k tokens
+	"qwen-turbo-latest":           1.4286,
+	"qwen-plus":                   1.4286,
+	"qwen-plus-latest":            1.4286,
+	"qwen-max":                    1.4286,
+	"qwen-max-latest":             1.4286,
+	"qwen-max-longcontext":        1.4286,
+	"qwen-vl-max":                 1.4286,
+	"qwen-vl-max-latest":          1.4286,
+	"qwen-vl-plus":                1.4286,
+	"qwen-vl-plus-latest":         1.4286,
+	"qwen-vl-ocr":                 1.4286,
+	"qwen-vl-ocr-latest":          1.4286,
+	"qwen-audio-turbo":            1.4286,
+	"qwen-math-plus":              1.4286,
+	"qwen-math-plus-latest":       1.4286,
+	"qwen-math-turbo":             1.4286,
+	"qwen-math-turbo-latest":      1.4286,
+	"qwen-coder-plus":             1.4286,
+	"qwen-coder-plus-latest":      1.4286,
+	"qwen-coder-turbo":            1.4286,
+	"qwen-coder-turbo-latest":     1.4286,
+	"qwq-32b-preview":             1.4286,
+	"qwen2.5-72b-instruct":        1.4286,
+	"qwen2.5-32b-instruct":        1.4286,
+	"qwen2.5-14b-instruct":        1.4286,
+	"qwen2.5-7b-instruct":         1.4286,
+	"qwen2.5-3b-instruct":         1.4286,
+	"qwen2.5-1.5b-instruct":       1.4286,
+	"qwen2.5-0.5b-instruct":       1.4286,
+	"qwen2-72b-instruct":          1.4286,
+	"qwen2-57b-a14b-instruct":     1.4286,
+	"qwen2-7b-instruct":           1.4286,
+	"qwen2-1.5b-instruct":         1.4286,
+	"qwen2-0.5b-instruct":         1.4286,
+	"qwen1.5-110b-chat":           1.4286,
+	"qwen1.5-72b-chat":            1.4286,
+	"qwen1.5-32b-chat":            1.4286,
+	"qwen1.5-14b-chat":            1.4286,
+	"qwen1.5-7b-chat":             1.4286,
+	"qwen1.5-1.8b-chat":           1.4286,
+	"qwen1.5-0.5b-chat":           1.4286,
+	"qwen-72b-chat":               1.4286,
+	"qwen-14b-chat":               1.4286,
+	"qwen-7b-chat":                1.4286,
+	"qwen-1.8b-chat":              1.4286,
+	"qwen-1.8b-longcontext-chat":  1.4286,
+	"qwen2-vl-7b-instruct":        1.4286,
+	"qwen2-vl-2b-instruct":        1.4286,
+	"qwen-vl-v1":                  1.4286,
+	"qwen-vl-chat-v1":             1.4286,
+	"qwen2-audio-instruct":        1.4286,
+	"qwen-audio-chat":             1.4286,
+	"qwen2.5-math-72b-instruct":   1.4286,
+	"qwen2.5-math-7b-instruct":    1.4286,
+	"qwen2.5-math-1.5b-instruct":  1.4286,
+	"qwen2-math-72b-instruct":     1.4286,
+	"qwen2-math-7b-instruct":      1.4286,
+	"qwen2-math-1.5b-instruct":    1.4286,
+	"qwen2.5-coder-32b-instruct":  1.4286,
+	"qwen2.5-coder-14b-instruct":  1.4286,
+	"qwen2.5-coder-7b-instruct":   1.4286,
+	"qwen2.5-coder-3b-instruct":   1.4286,
+	"qwen2.5-coder-1.5b-instruct": 1.4286,
+	"qwen2.5-coder-0.5b-instruct": 1.4286,
+	"text-embedding-v1":           0.05, // ￥0.0007 / 1k tokens
+	"text-embedding-v3":           0.05,
+	"text-embedding-v2":           0.05,
+	"text-embedding-async-v2":     0.05,
+	"text-embedding-async-v1":     0.05,
+	"ali-stable-diffusion-xl":     8.00,
+	"ali-stable-diffusion-v1.5":   8.00,
+	"wanx-v1":                     8.00,
+	"SparkDesk":                   1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v1.1":              1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v2.1":              1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.1":              1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.1-128K":         1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.5":              1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v3.5-32K":          1.2858, // ￥0.018 / 1k tokens
+	"SparkDesk-v4.0":              1.2858, // ￥0.018 / 1k tokens
+	"360GPT_S2_V9":                0.8572, // ¥0.012 / 1k tokens
+	"embedding-bert-512-v1":       0.0715, // ¥0.001 / 1k tokens
+	"embedding_s1_v1":             0.0715, // ¥0.001 / 1k tokens
+	"semantic_similarity_s1_v1":   0.0715, // ¥0.001 / 1k tokens
+	"hunyuan":                     7.143,  // ¥0.1 / 1k tokens  // https://cloud.tencent.com/document/product/1729/97731#e0e6be58-60c8-469f-bdeb-6c264ce3b4d0
+	"ChatStd":                     0.01 * RMB,
+	"ChatPro":                     0.1 * RMB,
+	// https://platform.moonshot.cn/pricing
+	"moonshot-v1-8k":   0.012 * RMB,
+	"moonshot-v1-32k":  0.024 * RMB,
+	"moonshot-v1-128k": 0.06 * RMB,
+	// https://platform.baichuan-ai.com/price
+	"Baichuan2-Turbo":      0.008 * RMB,
+	"Baichuan2-Turbo-192k": 0.016 * RMB,
+	"Baichuan2-53B":        0.02 * RMB,
+	// https://api.minimax.chat/document/price
+	"abab6.5-chat":  0.03 * RMB,
+	"abab6.5s-chat": 0.01 * RMB,
+	"abab6-chat":    0.1 * RMB,
+	"abab5.5-chat":  0.015 * RMB,
+	"abab5.5s-chat": 0.005 * RMB,
+	// https://docs.mistral.ai/platform/pricing/
+	"open-mistral-7b":       0.25 / 1000 * USD,
+	"open-mixtral-8x7b":     0.7 / 1000 * USD,
+	"mistral-small-latest":  2.0 / 1000 * USD,
+	"mistral-medium-latest": 2.7 / 1000 * USD,
+	"mistral-large-latest":  8.0 / 1000 * USD,
+	"mistral-embed":         0.1 / 1000 * USD,
+	// https://wow.groq.com/#:~:text=inquiries%C2%A0here.-,Model,-Current%20Speed
+	"gemma-7b-it":                           0.07 / 1000000 * USD,
+	"gemma2-9b-it":                          0.20 / 1000000 * USD,
+	"llama-3.1-70b-versatile":               0.59 / 1000000 * USD,
+	"llama-3.1-8b-instant":                  0.05 / 1000000 * USD,
+	"llama-3.2-11b-text-preview":            0.05 / 1000000 * USD,
+	"llama-3.2-11b-vision-preview":          0.05 / 1000000 * USD,
+	"llama-3.2-1b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-3b-preview":                  0.05 / 1000000 * USD,
+	"llama-3.2-90b-text-preview":            0.59 / 1000000 * USD,
+	"llama-guard-3-8b":                      0.05 / 1000000 * USD,
+	"llama3-70b-8192":                       0.59 / 1000000 * USD,
+	"llama3-8b-8192":                        0.05 / 1000000 * USD,
+	"llama3-groq-70b-8192-tool-use-preview": 0.89 / 1000000 * USD,
+	"llama3-groq-8b-8192-tool-use-preview":  0.19 / 1000000 * USD,
+	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
+
+	// https://platform.lingyiwanwu.com/docs#-计费单元
+	"yi-34b-chat-0205": 2.5 / 1000 * RMB,
+	"yi-34b-chat-200k": 12.0 / 1000 * RMB,
+	"yi-vl-plus":       6.0 / 1000 * RMB,
+	// https://platform.stepfun.com/docs/pricing/details
+	"step-1-8k":    0.005 / 1000 * RMB,
+	"step-1-32k":   0.015 / 1000 * RMB,
+	"step-1-128k":  0.040 / 1000 * RMB,
+	"step-1-256k":  0.095 / 1000 * RMB,
+	"step-1-flash": 0.001 / 1000 * RMB,
+	"step-2-16k":   0.038 / 1000 * RMB,
+	"step-1v-8k":   0.005 / 1000 * RMB,
+	"step-1v-32k":  0.015 / 1000 * RMB,
+	// aws llama3 https://aws.amazon.com/cn/bedrock/pricing/
+	"llama3-8b-8192(33)":  0.0003 / 0.002,  // $0.0003 / 1K tokens
+	"llama3-70b-8192(33)": 0.00265 / 0.002, // $0.00265 / 1K tokens
+	// https://cohere.com/pricing
+	"command":               0.5,
+	"command-nightly":       0.5,
+	"command-light":         0.5,
+	"command-light-nightly": 0.5,
+	"command-r":             0.5 / 1000 * USD,
+	"command-r-plus":        3.0 / 1000 * USD,
+	// https://platform.deepseek.com/api-docs/pricing/
+	"deepseek-chat":  1.0 / 1000 * RMB,
+	"deepseek-coder": 1.0 / 1000 * RMB,
+	// https://www.deepl.com/pro?cta=header-prices
+	"deepl-zh": 25.0 / 1000 * USD,
+	"deepl-en": 25.0 / 1000 * USD,
+	"deepl-ja": 25.0 / 1000 * USD,
+	// https://console.x.ai/
+	"grok-beta": 5.0 / 1000 * USD,
+	// replicate charges based on the number of generated images
+	// https://replicate.com/pricing
+	"black-forest-labs/flux-1.1-pro":                0.04 * USD,
+	"black-forest-labs/flux-1.1-pro-ultra":          0.06 * USD,
+	"black-forest-labs/flux-canny-dev":              0.025 * USD,
+	"black-forest-labs/flux-canny-pro":              0.05 * USD,
+	"black-forest-labs/flux-depth-dev":              0.025 * USD,
+	"black-forest-labs/flux-depth-pro":              0.05 * USD,
+	"black-forest-labs/flux-dev":                    0.025 * USD,
+	"black-forest-labs/flux-dev-lora":               0.032 * USD,
+	"black-forest-labs/flux-fill-dev":               0.04 * USD,
+	"black-forest-labs/flux-fill-pro":               0.05 * USD,
+	"black-forest-labs/flux-pro":                    0.055 * USD,
+	"black-forest-labs/flux-redux-dev":              0.025 * USD,
+	"black-forest-labs/flux-redux-schnell":          0.003 * USD,
+	"black-forest-labs/flux-schnell":                0.003 * USD,
+	"black-forest-labs/flux-schnell-lora":           0.02 * USD,
+	"ideogram-ai/ideogram-v2":                       0.08 * USD,
+	"ideogram-ai/ideogram-v2-turbo":                 0.05 * USD,
+	"recraft-ai/recraft-v3":                         0.04 * USD,
+	"recraft-ai/recraft-v3-svg":                     0.08 * USD,
+	"stability-ai/stable-diffusion-3":               0.035 * USD,
+	"stability-ai/stable-diffusion-3.5-large":       0.065 * USD,
+	"stability-ai/stable-diffusion-3.5-large-turbo": 0.04 * USD,
+	"stability-ai/stable-diffusion-3.5-medium":      0.035 * USD,
+	// replicate chat models
+	"ibm-granite/granite-20b-code-instruct-8k":  0.100 * USD,
+	"ibm-granite/granite-3.0-2b-instruct":       0.030 * USD,
+	"ibm-granite/granite-3.0-8b-instruct":       0.050 * USD,
+	"ibm-granite/granite-8b-code-instruct-128k": 0.050 * USD,
+	"meta/llama-2-13b":                          0.100 * USD,
+	"meta/llama-2-13b-chat":                     0.100 * USD,
+	"meta/llama-2-70b":                          0.650 * USD,
+	"meta/llama-2-70b-chat":                     0.650 * USD,
+	"meta/llama-2-7b":                           0.050 * USD,
+	"meta/llama-2-7b-chat":                      0.050 * USD,
+	"meta/meta-llama-3.1-405b-instruct":         9.500 * USD,
+	"meta/meta-llama-3-70b":                     0.650 * USD,
+	"meta/meta-llama-3-70b-instruct":            0.650 * USD,
+	"meta/meta-llama-3-8b":                      0.050 * USD,
+	"meta/meta-llama-3-8b-instruct":             0.050 * USD,
+	"mistralai/mistral-7b-instruct-v0.2":        0.050 * USD,
+	"mistralai/mistral-7b-v0.1":                 0.050 * USD,
+	"mistralai/mixtral-8x7b-instruct-v0.1":      0.300 * USD,
+}
+
+var CompletionRatio = map[string]float64{
+	// aws llama3
+	"llama3-8b-8192(33)":  0.0006 / 0.0003,
+	"llama3-70b-8192(33)": 0.0035 / 0.00265,
+}
+
+var (
+	DefaultModelRatio      map[string]float64
+	DefaultCompletionRatio map[string]float64
+)
+
+func init() {
+	DefaultModelRatio = make(map[string]float64)
+	for k, v := range ModelRatio {
+		DefaultModelRatio[k] = v
+	}
+	DefaultCompletionRatio = make(map[string]float64)
+	for k, v := range CompletionRatio {
+		DefaultCompletionRatio[k] = v
+	}
+}
+
+func AddNewMissingRatio(oldRatio string) string {
+	newRatio := make(map[string]float64)
+	err := json.Unmarshal([]byte(oldRatio), &newRatio)
+	if err != nil {
+		logger.SysError("error unmarshalling old ratio: " + err.Error())
+		return oldRatio
+	}
+	for k, v := range DefaultModelRatio {
+		if _, ok := newRatio[k]; !ok {
+			newRatio[k] = v
+		}
+	}
+	jsonBytes, err := json.Marshal(newRatio)
+	if err != nil {
+		logger.SysError("error marshalling new ratio: " + err.Error())
+		return oldRatio
+	}
+	return string(jsonBytes)
+}
+
+func ModelRatio2JSONString() string {
+	jsonBytes, err := json.Marshal(ModelRatio)
+	if err != nil {
+		logger.SysError("error marshalling model ratio: " + err.Error())
+	}
+	return string(jsonBytes)
+}
+
+func UpdateModelRatioByJSONString(jsonStr string) error {
+	ModelRatio = make(map[string]float64)
+	return json.Unmarshal([]byte(jsonStr), &ModelRatio)
+}
+
+func GetModelRatio(name string, channelType int) float64 {
+	if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
+		name = strings.TrimSuffix(name, "-internet")
+	}
+	if strings.HasPrefix(name, "command-") && strings.HasSuffix(name, "-internet") {
+		name = strings.TrimSuffix(name, "-internet")
+	}
+	model := fmt.Sprintf("%s(%d)", name, channelType)
+	if ratio, ok := ModelRatio[model]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultModelRatio[model]; ok {
+		return ratio
+	}
+	if ratio, ok := ModelRatio[name]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultModelRatio[name]; ok {
+		return ratio
+	}
+	logger.SysError("model ratio not found: " + name)
+	return 30
+}
+
+func CompletionRatio2JSONString() string {
+	jsonBytes, err := json.Marshal(CompletionRatio)
+	if err != nil {
+		logger.SysError("error marshalling completion ratio: " + err.Error())
+	}
+	return string(jsonBytes)
+}
+
+func UpdateCompletionRatioByJSONString(jsonStr string) error {
+	CompletionRatio = make(map[string]float64)
+	return json.Unmarshal([]byte(jsonStr), &CompletionRatio)
+}
+
+func GetCompletionRatio(name string, channelType int) float64 {
+	if strings.HasPrefix(name, "qwen-") && strings.HasSuffix(name, "-internet") {
+		name = strings.TrimSuffix(name, "-internet")
+	}
+	model := fmt.Sprintf("%s(%d)", name, channelType)
+	if ratio, ok := CompletionRatio[model]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultCompletionRatio[model]; ok {
+		return ratio
+	}
+	if ratio, ok := CompletionRatio[name]; ok {
+		return ratio
+	}
+	if ratio, ok := DefaultCompletionRatio[name]; ok {
+		return ratio
+	}
+	if strings.HasPrefix(name, "gpt-3.5") {
+		if name == "gpt-3.5-turbo" || strings.HasSuffix(name, "0125") {
+			// https://openai.com/blog/new-embedding-models-and-api-updates
+			// Updated GPT-3.5 Turbo model and lower pricing
+			return 3
+		}
+		if strings.HasSuffix(name, "1106") {
+			return 2
+		}
+		return 4.0 / 3.0
+	}
+	if strings.HasPrefix(name, "gpt-4") {
+		if strings.HasPrefix(name, "gpt-4o") {
+			if name == "gpt-4o-2024-05-13" {
+				return 3
+			}
+			return 4
+		}
+		if strings.HasPrefix(name, "gpt-4-turbo") ||
+			strings.HasSuffix(name, "preview") {
+			return 3
+		}
+		return 2
+	}
+	// including o1, o1-preview, o1-mini
+	if strings.HasPrefix(name, "o1") {
+		return 4
+	}
+	if name == "chatgpt-4o-latest" {
+		return 3
+	}
+	if strings.HasPrefix(name, "claude-3") {
+		return 5
+	}
+	if strings.HasPrefix(name, "claude-") {
+		return 3
+	}
+	if strings.HasPrefix(name, "mistral-") {
+		return 3
+	}
+	if strings.HasPrefix(name, "gemini-") {
+		return 3
+	}
+	if strings.HasPrefix(name, "deepseek-") {
+		return 2
+	}
+
+	switch name {
+	case "llama2-70b-4096":
+		return 0.8 / 0.64
+	case "llama3-8b-8192":
+		return 2
+	case "llama3-70b-8192":
+		return 0.79 / 0.59
+	case "command", "command-light", "command-nightly", "command-light-nightly":
+		return 2
+	case "command-r":
+		return 3
+	case "command-r-plus":
+		return 5
+	case "grok-beta":
+		return 3
+	// Replicate Models
+	// https://replicate.com/pricing
+	case "ibm-granite/granite-20b-code-instruct-8k":
+		return 5
+	case "ibm-granite/granite-3.0-2b-instruct":
+		return 8.333333333333334
+	case "ibm-granite/granite-3.0-8b-instruct",
+		"ibm-granite/granite-8b-code-instruct-128k":
+		return 5
+	case "meta/llama-2-13b",
+		"meta/llama-2-13b-chat",
+		"meta/llama-2-7b",
+		"meta/llama-2-7b-chat",
+		"meta/meta-llama-3-8b",
+		"meta/meta-llama-3-8b-instruct":
+		return 5
+	case "meta/llama-2-70b",
+		"meta/llama-2-70b-chat",
+		"meta/meta-llama-3-70b",
+		"meta/meta-llama-3-70b-instruct":
+		return 2.750 / 0.650 // ≈4.230769
+	case "meta/meta-llama-3.1-405b-instruct":
+		return 1
+	case "mistralai/mistral-7b-instruct-v0.2",
+		"mistralai/mistral-7b-v0.1":
+		return 5
+	case "mistralai/mixtral-8x7b-instruct-v0.1":
+		return 1.000 / 0.300 // ≈3.333333
+	}
+
+	return 1
+}
diff --git a/relay/channel/ali/adaptor.go b/relay/channel/ali/adaptor.go
deleted file mode 100644
index 6a3245ad..00000000
--- a/relay/channel/ali/adaptor.go
+++ /dev/null
@@ -1,86 +0,0 @@
-package ali
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-)
-
-// https://help.aliyun.com/zh/dashscope/developer-reference/api-details
-
-type Adaptor struct {
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	fullRequestURL := fmt.Sprintf("%s/api/v1/services/aigc/text-generation/generation", meta.BaseURL)
-	if meta.Mode == constant.RelayModeEmbeddings {
-		fullRequestURL = fmt.Sprintf("%s/api/v1/services/embeddings/text-embedding/text-embedding", meta.BaseURL)
-	}
-	return fullRequestURL, nil
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
-	if meta.IsStream {
-		req.Header.Set("Accept", "text/event-stream")
-	}
-	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
-	if meta.IsStream {
-		req.Header.Set("X-DashScope-SSE", "enable")
-	}
-	if c.GetString(common.ConfigKeyPlugin) != "" {
-		req.Header.Set("X-DashScope-Plugin", c.GetString(common.ConfigKeyPlugin))
-	}
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	switch relayMode {
-	case constant.RelayModeEmbeddings:
-		baiduEmbeddingRequest := ConvertEmbeddingRequest(*request)
-		return baiduEmbeddingRequest, nil
-	default:
-		baiduRequest := ConvertRequest(*request)
-		return baiduRequest, nil
-	}
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		err, usage = StreamHandler(c, resp)
-	} else {
-		switch meta.Mode {
-		case constant.RelayModeEmbeddings:
-			err, usage = EmbeddingHandler(c, resp)
-		default:
-			err, usage = Handler(c, resp)
-		}
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	return ModelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "ali"
-}
diff --git a/relay/channel/ali/constants.go b/relay/channel/ali/constants.go
deleted file mode 100644
index 16bcfca4..00000000
--- a/relay/channel/ali/constants.go
+++ /dev/null
@@ -1,6 +0,0 @@
-package ali
-
-var ModelList = []string{
-	"qwen-turbo", "qwen-plus", "qwen-max", "qwen-max-longcontext",
-	"text-embedding-v1",
-}
diff --git a/relay/channel/ali/model.go b/relay/channel/ali/model.go
deleted file mode 100644
index 76e814d1..00000000
--- a/relay/channel/ali/model.go
+++ /dev/null
@@ -1,73 +0,0 @@
-package ali
-
-type Message struct {
-	Content string `json:"content"`
-	Role    string `json:"role"`
-}
-
-type Input struct {
-	//Prompt   string       `json:"prompt"`
-	Messages []Message `json:"messages"`
-}
-
-type Parameters struct {
-	TopP              float64 `json:"top_p,omitempty"`
-	TopK              int     `json:"top_k,omitempty"`
-	Seed              uint64  `json:"seed,omitempty"`
-	EnableSearch      bool    `json:"enable_search,omitempty"`
-	IncrementalOutput bool    `json:"incremental_output,omitempty"`
-	MaxTokens         int     `json:"max_tokens,omitempty"`
-	Temperature       float64 `json:"temperature,omitempty"`
-}
-
-type ChatRequest struct {
-	Model      string     `json:"model"`
-	Input      Input      `json:"input"`
-	Parameters Parameters `json:"parameters,omitempty"`
-}
-
-type EmbeddingRequest struct {
-	Model string `json:"model"`
-	Input struct {
-		Texts []string `json:"texts"`
-	} `json:"input"`
-	Parameters *struct {
-		TextType string `json:"text_type,omitempty"`
-	} `json:"parameters,omitempty"`
-}
-
-type Embedding struct {
-	Embedding []float64 `json:"embedding"`
-	TextIndex int       `json:"text_index"`
-}
-
-type EmbeddingResponse struct {
-	Output struct {
-		Embeddings []Embedding `json:"embeddings"`
-	} `json:"output"`
-	Usage Usage `json:"usage"`
-	Error
-}
-
-type Error struct {
-	Code      string `json:"code"`
-	Message   string `json:"message"`
-	RequestId string `json:"request_id"`
-}
-
-type Usage struct {
-	InputTokens  int `json:"input_tokens"`
-	OutputTokens int `json:"output_tokens"`
-	TotalTokens  int `json:"total_tokens"`
-}
-
-type Output struct {
-	Text         string `json:"text"`
-	FinishReason string `json:"finish_reason"`
-}
-
-type ChatResponse struct {
-	Output Output `json:"output"`
-	Usage  Usage  `json:"usage"`
-	Error
-}
diff --git a/relay/channel/baidu/constants.go b/relay/channel/baidu/constants.go
deleted file mode 100644
index 45a4e901..00000000
--- a/relay/channel/baidu/constants.go
+++ /dev/null
@@ -1,13 +0,0 @@
-package baidu
-
-var ModelList = []string{
-	"ERNIE-Bot-4",
-	"ERNIE-Bot-8K",
-	"ERNIE-Bot",
-	"ERNIE-Speed",
-	"ERNIE-Bot-turbo",
-	"Embedding-V1",
-	"bge-large-zh",
-	"bge-large-en",
-	"tao-8k",
-}
diff --git a/relay/channel/gemini/adaptor.go b/relay/channel/gemini/adaptor.go
deleted file mode 100644
index f3305e5d..00000000
--- a/relay/channel/gemini/adaptor.go
+++ /dev/null
@@ -1,66 +0,0 @@
-package gemini
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common/helper"
-	channelhelper "github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-)
-
-type Adaptor struct {
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	version := helper.AssignOrDefault(meta.APIVersion, "v1")
-	action := "generateContent"
-	if meta.IsStream {
-		action = "streamGenerateContent"
-	}
-	return fmt.Sprintf("%s/%s/models/%s:%s", meta.BaseURL, version, meta.ActualModelName, action), nil
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channelhelper.SetupCommonRequestHeader(c, req, meta)
-	req.Header.Set("x-goog-api-key", meta.APIKey)
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	return ConvertRequest(*request), nil
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channelhelper.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		var responseText string
-		err, responseText = StreamHandler(c, resp)
-		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
-	} else {
-		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	return ModelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "google gemini"
-}
diff --git a/relay/channel/gemini/constants.go b/relay/channel/gemini/constants.go
deleted file mode 100644
index e8d3a155..00000000
--- a/relay/channel/gemini/constants.go
+++ /dev/null
@@ -1,8 +0,0 @@
-package gemini
-
-// https://ai.google.dev/models/gemini
-
-var ModelList = []string{
-	"gemini-pro", "gemini-1.0-pro-001",
-	"gemini-pro-vision", "gemini-1.0-pro-vision-001",
-}
diff --git a/relay/channel/gemini/model.go b/relay/channel/gemini/model.go
deleted file mode 100644
index d1e3c4fd..00000000
--- a/relay/channel/gemini/model.go
+++ /dev/null
@@ -1,41 +0,0 @@
-package gemini
-
-type ChatRequest struct {
-	Contents         []ChatContent        `json:"contents"`
-	SafetySettings   []ChatSafetySettings `json:"safety_settings,omitempty"`
-	GenerationConfig ChatGenerationConfig `json:"generation_config,omitempty"`
-	Tools            []ChatTools          `json:"tools,omitempty"`
-}
-
-type InlineData struct {
-	MimeType string `json:"mimeType"`
-	Data     string `json:"data"`
-}
-
-type Part struct {
-	Text       string      `json:"text,omitempty"`
-	InlineData *InlineData `json:"inlineData,omitempty"`
-}
-
-type ChatContent struct {
-	Role  string `json:"role,omitempty"`
-	Parts []Part `json:"parts"`
-}
-
-type ChatSafetySettings struct {
-	Category  string `json:"category"`
-	Threshold string `json:"threshold"`
-}
-
-type ChatTools struct {
-	FunctionDeclarations any `json:"functionDeclarations,omitempty"`
-}
-
-type ChatGenerationConfig struct {
-	Temperature     float64  `json:"temperature,omitempty"`
-	TopP            float64  `json:"topP,omitempty"`
-	TopK            float64  `json:"topK,omitempty"`
-	MaxOutputTokens int      `json:"maxOutputTokens,omitempty"`
-	CandidateCount  int      `json:"candidateCount,omitempty"`
-	StopSequences   []string `json:"stopSequences,omitempty"`
-}
diff --git a/relay/channel/groq/constants.go b/relay/channel/groq/constants.go
deleted file mode 100644
index fc9a9ebd..00000000
--- a/relay/channel/groq/constants.go
+++ /dev/null
@@ -1,10 +0,0 @@
-package groq
-
-// https://console.groq.com/docs/models
-
-var ModelList = []string{
-	"gemma-7b-it",
-	"llama2-7b-2048",
-	"llama2-70b-4096",
-	"mixtral-8x7b-32768",
-}
diff --git a/relay/channel/interface.go b/relay/channel/interface.go
deleted file mode 100644
index e25db83f..00000000
--- a/relay/channel/interface.go
+++ /dev/null
@@ -1,20 +0,0 @@
-package channel
-
-import (
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-)
-
-type Adaptor interface {
-	Init(meta *util.RelayMeta)
-	GetRequestURL(meta *util.RelayMeta) (string, error)
-	SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error
-	ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error)
-	DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error)
-	DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode)
-	GetModelList() []string
-	GetChannelName() string
-}
diff --git a/relay/channel/minimax/constants.go b/relay/channel/minimax/constants.go
deleted file mode 100644
index c3da5b2d..00000000
--- a/relay/channel/minimax/constants.go
+++ /dev/null
@@ -1,7 +0,0 @@
-package minimax
-
-var ModelList = []string{
-	"abab5.5s-chat",
-	"abab5.5-chat",
-	"abab6-chat",
-}
diff --git a/relay/channel/minimax/main.go b/relay/channel/minimax/main.go
deleted file mode 100644
index a01821c2..00000000
--- a/relay/channel/minimax/main.go
+++ /dev/null
@@ -1,14 +0,0 @@
-package minimax
-
-import (
-	"fmt"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/util"
-)
-
-func GetRequestURL(meta *util.RelayMeta) (string, error) {
-	if meta.Mode == constant.RelayModeChatCompletions {
-		return fmt.Sprintf("%s/v1/text/chatcompletion_v2", meta.BaseURL), nil
-	}
-	return "", fmt.Errorf("unsupported relay mode %d for minimax", meta.Mode)
-}
diff --git a/relay/channel/ollama/adaptor.go b/relay/channel/ollama/adaptor.go
deleted file mode 100644
index 06c66101..00000000
--- a/relay/channel/ollama/adaptor.go
+++ /dev/null
@@ -1,65 +0,0 @@
-package ollama
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-)
-
-type Adaptor struct {
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	// https://github.com/ollama/ollama/blob/main/docs/api.md
-	fullRequestURL := fmt.Sprintf("%s/api/chat", meta.BaseURL)
-	return fullRequestURL, nil
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
-	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	switch relayMode {
-	case constant.RelayModeEmbeddings:
-		return nil, errors.New("not supported")
-	default:
-		return ConvertRequest(*request), nil
-	}
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		err, usage = StreamHandler(c, resp)
-	} else {
-		err, usage = Handler(c, resp)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	return ModelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "ollama"
-}
diff --git a/relay/channel/ollama/constants.go b/relay/channel/ollama/constants.go
deleted file mode 100644
index 32f82b2a..00000000
--- a/relay/channel/ollama/constants.go
+++ /dev/null
@@ -1,5 +0,0 @@
-package ollama
-
-var ModelList = []string{
-	"qwen:0.5b-chat",
-}
diff --git a/relay/channel/openai/adaptor.go b/relay/channel/openai/adaptor.go
deleted file mode 100644
index 47594030..00000000
--- a/relay/channel/openai/adaptor.go
+++ /dev/null
@@ -1,92 +0,0 @@
-package openai
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/minimax"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-	"strings"
-)
-
-type Adaptor struct {
-	ChannelType int
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-	a.ChannelType = meta.ChannelType
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	switch meta.ChannelType {
-	case common.ChannelTypeAzure:
-		// https://learn.microsoft.com/en-us/azure/cognitive-services/openai/chatgpt-quickstart?pivots=rest-api&tabs=command-line#rest-api
-		requestURL := strings.Split(meta.RequestURLPath, "?")[0]
-		requestURL = fmt.Sprintf("%s?api-version=%s", requestURL, meta.APIVersion)
-		task := strings.TrimPrefix(requestURL, "/v1/")
-		model_ := meta.ActualModelName
-		model_ = strings.Replace(model_, ".", "", -1)
-		// https://github.com/songquanpeng/one-api/issues/67
-		model_ = strings.TrimSuffix(model_, "-0301")
-		model_ = strings.TrimSuffix(model_, "-0314")
-		model_ = strings.TrimSuffix(model_, "-0613")
-
-		requestURL = fmt.Sprintf("/openai/deployments/%s/%s", model_, task)
-		return util.GetFullRequestURL(meta.BaseURL, requestURL, meta.ChannelType), nil
-	case common.ChannelTypeMinimax:
-		return minimax.GetRequestURL(meta)
-	default:
-		return util.GetFullRequestURL(meta.BaseURL, meta.RequestURLPath, meta.ChannelType), nil
-	}
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
-	if meta.ChannelType == common.ChannelTypeAzure {
-		req.Header.Set("api-key", meta.APIKey)
-		return nil
-	}
-	req.Header.Set("Authorization", "Bearer "+meta.APIKey)
-	if meta.ChannelType == common.ChannelTypeOpenRouter {
-		req.Header.Set("HTTP-Referer", "https://github.com/songquanpeng/one-api")
-		req.Header.Set("X-Title", "One API")
-	}
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	return request, nil
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		var responseText string
-		err, responseText, _ = StreamHandler(c, resp, meta.Mode)
-		usage = ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
-	} else {
-		err, usage = Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	_, modelList := GetCompatibleChannelMeta(a.ChannelType)
-	return modelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	channelName, _ := GetCompatibleChannelMeta(a.ChannelType)
-	return channelName
-}
diff --git a/relay/channel/openai/compatible.go b/relay/channel/openai/compatible.go
deleted file mode 100644
index e4951a34..00000000
--- a/relay/channel/openai/compatible.go
+++ /dev/null
@@ -1,46 +0,0 @@
-package openai
-
-import (
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/channel/ai360"
-	"github.com/songquanpeng/one-api/relay/channel/baichuan"
-	"github.com/songquanpeng/one-api/relay/channel/groq"
-	"github.com/songquanpeng/one-api/relay/channel/lingyiwanwu"
-	"github.com/songquanpeng/one-api/relay/channel/minimax"
-	"github.com/songquanpeng/one-api/relay/channel/mistral"
-	"github.com/songquanpeng/one-api/relay/channel/moonshot"
-)
-
-var CompatibleChannels = []int{
-	common.ChannelTypeAzure,
-	common.ChannelType360,
-	common.ChannelTypeMoonshot,
-	common.ChannelTypeBaichuan,
-	common.ChannelTypeMinimax,
-	common.ChannelTypeMistral,
-	common.ChannelTypeGroq,
-	common.ChannelTypeLingYiWanWu,
-}
-
-func GetCompatibleChannelMeta(channelType int) (string, []string) {
-	switch channelType {
-	case common.ChannelTypeAzure:
-		return "azure", ModelList
-	case common.ChannelType360:
-		return "360", ai360.ModelList
-	case common.ChannelTypeMoonshot:
-		return "moonshot", moonshot.ModelList
-	case common.ChannelTypeBaichuan:
-		return "baichuan", baichuan.ModelList
-	case common.ChannelTypeMinimax:
-		return "minimax", minimax.ModelList
-	case common.ChannelTypeMistral:
-		return "mistralai", mistral.ModelList
-	case common.ChannelTypeGroq:
-		return "groq", groq.ModelList
-	case common.ChannelTypeLingYiWanWu:
-		return "lingyiwanwu", lingyiwanwu.ModelList
-	default:
-		return "openai", ModelList
-	}
-}
diff --git a/relay/channel/openai/helper.go b/relay/channel/openai/helper.go
deleted file mode 100644
index 9bca8cab..00000000
--- a/relay/channel/openai/helper.go
+++ /dev/null
@@ -1,11 +0,0 @@
-package openai
-
-import "github.com/songquanpeng/one-api/relay/model"
-
-func ResponseText2Usage(responseText string, modeName string, promptTokens int) *model.Usage {
-	usage := &model.Usage{}
-	usage.PromptTokens = promptTokens
-	usage.CompletionTokens = CountTokenText(responseText, modeName)
-	usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
-	return usage
-}
diff --git a/relay/channel/tencent/adaptor.go b/relay/channel/tencent/adaptor.go
deleted file mode 100644
index f348674e..00000000
--- a/relay/channel/tencent/adaptor.go
+++ /dev/null
@@ -1,76 +0,0 @@
-package tencent
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"net/http"
-	"strings"
-)
-
-// https://cloud.tencent.com/document/api/1729/101837
-
-type Adaptor struct {
-	Sign string
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	return fmt.Sprintf("%s/hyllm/v1/chat/completions", meta.BaseURL), nil
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
-	req.Header.Set("Authorization", a.Sign)
-	req.Header.Set("X-TC-Action", meta.ActualModelName)
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	apiKey := c.Request.Header.Get("Authorization")
-	apiKey = strings.TrimPrefix(apiKey, "Bearer ")
-	appId, secretId, secretKey, err := ParseConfig(apiKey)
-	if err != nil {
-		return nil, err
-	}
-	tencentRequest := ConvertRequest(*request)
-	tencentRequest.AppId = appId
-	tencentRequest.SecretId = secretId
-	// we have to calculate the sign here
-	a.Sign = GetSign(*tencentRequest, secretKey)
-	return tencentRequest, nil
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		var responseText string
-		err, responseText = StreamHandler(c, resp)
-		usage = openai.ResponseText2Usage(responseText, meta.ActualModelName, meta.PromptTokens)
-	} else {
-		err, usage = Handler(c, resp)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	return ModelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "tencent"
-}
diff --git a/relay/channel/tencent/constants.go b/relay/channel/tencent/constants.go
deleted file mode 100644
index fe176c2c..00000000
--- a/relay/channel/tencent/constants.go
+++ /dev/null
@@ -1,7 +0,0 @@
-package tencent
-
-var ModelList = []string{
-	"ChatPro",
-	"ChatStd",
-	"hunyuan",
-}
diff --git a/relay/channel/tencent/model.go b/relay/channel/tencent/model.go
deleted file mode 100644
index 71286be9..00000000
--- a/relay/channel/tencent/model.go
+++ /dev/null
@@ -1,63 +0,0 @@
-package tencent
-
-import (
-	"github.com/songquanpeng/one-api/relay/model"
-)
-
-type Message struct {
-	Role    string `json:"role"`
-	Content string `json:"content"`
-}
-
-type ChatRequest struct {
-	AppId    int64  `json:"app_id"`    // 腾讯云账号的 APPID
-	SecretId string `json:"secret_id"` // 官网 SecretId
-	// Timestamp当前 UNIX 时间戳，单位为秒，可记录发起 API 请求的时间。
-	// 例如1529223702，如果与当前时间相差过大，会引起签名过期错误
-	Timestamp int64 `json:"timestamp"`
-	// Expired 签名的有效期，是一个符合 UNIX Epoch 时间戳规范的数值，
-	// 单位为秒；Expired 必须大于 Timestamp 且 Expired-Timestamp 小于90天
-	Expired int64  `json:"expired"`
-	QueryID string `json:"query_id"` //请求 Id，用于问题排查
-	// Temperature 较高的数值会使输出更加随机，而较低的数值会使其更加集中和确定
-	// 默认 1.0，取值区间为[0.0,2.0]，非必要不建议使用,不合理的取值会影响效果
-	// 建议该参数和 top_p 只设置1个，不要同时更改 top_p
-	Temperature float64 `json:"temperature"`
-	// TopP 影响输出文本的多样性，取值越大，生成文本的多样性越强
-	// 默认1.0，取值区间为[0.0, 1.0]，非必要不建议使用, 不合理的取值会影响效果
-	// 建议该参数和 temperature 只设置1个，不要同时更改
-	TopP float64 `json:"top_p"`
-	// Stream 0：同步，1：流式 （默认，协议：SSE)
-	// 同步请求超时：60s，如果内容较长建议使用流式
-	Stream int `json:"stream"`
-	// Messages 会话内容, 长度最多为40, 按对话时间从旧到新在数组中排列
-	// 输入 content 总数最大支持 3000 token。
-	Messages []Message `json:"messages"`
-}
-
-type Error struct {
-	Code    int    `json:"code"`
-	Message string `json:"message"`
-}
-
-type Usage struct {
-	InputTokens  int `json:"input_tokens"`
-	OutputTokens int `json:"output_tokens"`
-	TotalTokens  int `json:"total_tokens"`
-}
-
-type ResponseChoices struct {
-	FinishReason string  `json:"finish_reason,omitempty"` // 流式结束标志位，为 stop 则表示尾包
-	Messages     Message `json:"messages,omitempty"`      // 内容，同步模式返回内容，流模式为 null 输出 content 内容总数最多支持 1024token。
-	Delta        Message `json:"delta,omitempty"`         // 内容，流模式返回内容，同步模式为 null 输出 content 内容总数最多支持 1024token。
-}
-
-type ChatResponse struct {
-	Choices []ResponseChoices `json:"choices,omitempty"` // 结果
-	Created string            `json:"created,omitempty"` // unix 时间戳的字符串
-	Id      string            `json:"id,omitempty"`      // 会话 id
-	Usage   model.Usage       `json:"usage,omitempty"`   // token 数量
-	Error   Error             `json:"error,omitempty"`   // 错误信息 注意：此字段可能返回 null，表示取不到有效值
-	Note    string            `json:"note,omitempty"`    // 注释
-	ReqID   string            `json:"req_id,omitempty"`  // 唯一请求 Id，每次请求都会返回。用于反馈接口入参
-}
diff --git a/relay/channel/zhipu/adaptor.go b/relay/channel/zhipu/adaptor.go
deleted file mode 100644
index 0ca23d59..00000000
--- a/relay/channel/zhipu/adaptor.go
+++ /dev/null
@@ -1,101 +0,0 @@
-package zhipu
-
-import (
-	"errors"
-	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"io"
-	"math"
-	"net/http"
-	"strings"
-)
-
-type Adaptor struct {
-	APIVersion string
-}
-
-func (a *Adaptor) Init(meta *util.RelayMeta) {
-
-}
-
-func (a *Adaptor) SetVersionByModeName(modelName string) {
-	if strings.HasPrefix(modelName, "glm-") {
-		a.APIVersion = "v4"
-	} else {
-		a.APIVersion = "v3"
-	}
-}
-
-func (a *Adaptor) GetRequestURL(meta *util.RelayMeta) (string, error) {
-	a.SetVersionByModeName(meta.ActualModelName)
-	if a.APIVersion == "v4" {
-		return fmt.Sprintf("%s/api/paas/v4/chat/completions", meta.BaseURL), nil
-	}
-	method := "invoke"
-	if meta.IsStream {
-		method = "sse-invoke"
-	}
-	return fmt.Sprintf("%s/api/paas/v3/model-api/%s/%s", meta.BaseURL, meta.ActualModelName, method), nil
-}
-
-func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Request, meta *util.RelayMeta) error {
-	channel.SetupCommonRequestHeader(c, req, meta)
-	token := GetToken(meta.APIKey)
-	req.Header.Set("Authorization", token)
-	return nil
-}
-
-func (a *Adaptor) ConvertRequest(c *gin.Context, relayMode int, request *model.GeneralOpenAIRequest) (any, error) {
-	if request == nil {
-		return nil, errors.New("request is nil")
-	}
-	// TopP (0.0, 1.0)
-	request.TopP = math.Min(0.99, request.TopP)
-	request.TopP = math.Max(0.01, request.TopP)
-
-	// Temperature (0.0, 1.0)
-	request.Temperature = math.Min(0.99, request.Temperature)
-	request.Temperature = math.Max(0.01, request.Temperature)
-	a.SetVersionByModeName(request.Model)
-	if a.APIVersion == "v4" {
-		return request, nil
-	}
-	return ConvertRequest(*request), nil
-}
-
-func (a *Adaptor) DoRequest(c *gin.Context, meta *util.RelayMeta, requestBody io.Reader) (*http.Response, error) {
-	return channel.DoRequestHelper(a, c, meta, requestBody)
-}
-
-func (a *Adaptor) DoResponseV4(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if meta.IsStream {
-		err, _, usage = openai.StreamHandler(c, resp, meta.Mode)
-	} else {
-		err, usage = openai.Handler(c, resp, meta.PromptTokens, meta.ActualModelName)
-	}
-	return
-}
-
-func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, meta *util.RelayMeta) (usage *model.Usage, err *model.ErrorWithStatusCode) {
-	if a.APIVersion == "v4" {
-		return a.DoResponseV4(c, resp, meta)
-	}
-	if meta.IsStream {
-		err, usage = StreamHandler(c, resp)
-	} else {
-		err, usage = Handler(c, resp)
-	}
-	return
-}
-
-func (a *Adaptor) GetModelList() []string {
-	return ModelList
-}
-
-func (a *Adaptor) GetChannelName() string {
-	return "zhipu"
-}
diff --git a/relay/channeltype/define.go b/relay/channeltype/define.go
new file mode 100644
index 00000000..f54d0e30
--- /dev/null
+++ b/relay/channeltype/define.go
@@ -0,0 +1,52 @@
+package channeltype
+
+const (
+	Unknown = iota
+	OpenAI
+	API2D
+	Azure
+	CloseAI
+	OpenAISB
+	OpenAIMax
+	OhMyGPT
+	Custom
+	Ails
+	AIProxy
+	PaLM
+	API2GPT
+	AIGC2D
+	Anthropic
+	Baidu
+	Zhipu
+	Ali
+	Xunfei
+	AI360
+	OpenRouter
+	AIProxyLibrary
+	FastGPT
+	Tencent
+	Gemini
+	Moonshot
+	Baichuan
+	Minimax
+	Mistral
+	Groq
+	Ollama
+	LingYiWanWu
+	StepFun
+	AwsClaude
+	Coze
+	Cohere
+	DeepSeek
+	Cloudflare
+	DeepL
+	TogetherAI
+	Doubao
+	Novita
+	VertextAI
+	Proxy
+	SiliconFlow
+	XAI
+	Replicate
+	Dummy
+)
diff --git a/relay/channeltype/helper.go b/relay/channeltype/helper.go
new file mode 100644
index 00000000..8839b30a
--- /dev/null
+++ b/relay/channeltype/helper.go
@@ -0,0 +1,47 @@
+package channeltype
+
+import "github.com/songquanpeng/one-api/relay/apitype"
+
+func ToAPIType(channelType int) int {
+	apiType := apitype.OpenAI
+	switch channelType {
+	case Anthropic:
+		apiType = apitype.Anthropic
+	case Baidu:
+		apiType = apitype.Baidu
+	case PaLM:
+		apiType = apitype.PaLM
+	case Zhipu:
+		apiType = apitype.Zhipu
+	case Ali:
+		apiType = apitype.Ali
+	case Xunfei:
+		apiType = apitype.Xunfei
+	case AIProxyLibrary:
+		apiType = apitype.AIProxyLibrary
+	case Tencent:
+		apiType = apitype.Tencent
+	case Gemini:
+		apiType = apitype.Gemini
+	case Ollama:
+		apiType = apitype.Ollama
+	case AwsClaude:
+		apiType = apitype.AwsClaude
+	case Coze:
+		apiType = apitype.Coze
+	case Cohere:
+		apiType = apitype.Cohere
+	case Cloudflare:
+		apiType = apitype.Cloudflare
+	case DeepL:
+		apiType = apitype.DeepL
+	case VertextAI:
+		apiType = apitype.VertexAI
+	case Replicate:
+		apiType = apitype.Replicate
+	case Proxy:
+		apiType = apitype.Proxy
+	}
+
+	return apiType
+}
diff --git a/relay/channeltype/url.go b/relay/channeltype/url.go
new file mode 100644
index 00000000..8e271f4e
--- /dev/null
+++ b/relay/channeltype/url.go
@@ -0,0 +1,57 @@
+package channeltype
+
+var ChannelBaseURLs = []string{
+	"",                              // 0
+	"https://api.openai.com",        // 1
+	"https://oa.api2d.net",          // 2
+	"",                              // 3
+	"https://api.closeai-proxy.xyz", // 4
+	"https://api.openai-sb.com",     // 5
+	"https://api.openaimax.com",     // 6
+	"https://api.ohmygpt.com",       // 7
+	"",                              // 8
+	"https://api.caipacity.com",     // 9
+	"https://api.aiproxy.io",        // 10
+	"https://generativelanguage.googleapis.com", // 11
+	"https://api.api2gpt.com",                   // 12
+	"https://api.aigc2d.com",                    // 13
+	"https://api.anthropic.com",                 // 14
+	"https://aip.baidubce.com",                  // 15
+	"https://open.bigmodel.cn",                  // 16
+	"https://dashscope.aliyuncs.com",            // 17
+	"",                                          // 18
+	"https://ai.360.cn",                         // 19
+	"https://openrouter.ai/api",                 // 20
+	"https://api.aiproxy.io",                    // 21
+	"https://fastgpt.run/api/openapi",           // 22
+	"https://hunyuan.tencentcloudapi.com",       // 23
+	"https://generativelanguage.googleapis.com", // 24
+	"https://api.moonshot.cn",                   // 25
+	"https://api.baichuan-ai.com",               // 26
+	"https://api.minimax.chat",                  // 27
+	"https://api.mistral.ai",                    // 28
+	"https://api.groq.com/openai",               // 29
+	"http://localhost:11434",                    // 30
+	"https://api.lingyiwanwu.com",               // 31
+	"https://api.stepfun.com",                   // 32
+	"",                                          // 33
+	"https://api.coze.com",                      // 34
+	"https://api.cohere.ai",                     // 35
+	"https://api.deepseek.com",                  // 36
+	"https://api.cloudflare.com",                // 37
+	"https://api-free.deepl.com",                // 38
+	"https://api.together.xyz",                  // 39
+	"https://ark.cn-beijing.volces.com",         // 40
+	"https://api.novita.ai/v3/openai",           // 41
+	"",                                          // 42
+	"",                                          // 43
+	"https://api.siliconflow.cn",                // 44
+	"https://api.x.ai",                          // 45
+	"https://api.replicate.com/v1/models/",      // 46
+}
+
+func init() {
+	if len(ChannelBaseURLs) != Dummy {
+		panic("channel base urls length not match")
+	}
+}
diff --git a/relay/channeltype/url_test.go b/relay/channeltype/url_test.go
new file mode 100644
index 00000000..2aada279
--- /dev/null
+++ b/relay/channeltype/url_test.go
@@ -0,0 +1,12 @@
+package channeltype
+
+import (
+	. "github.com/smartystreets/goconvey/convey"
+	"testing"
+)
+
+func TestChannelBaseURLs(t *testing.T) {
+	Convey("channel base urls", t, func() {
+		So(len(ChannelBaseURLs), ShouldEqual, Dummy)
+	})
+}
diff --git a/relay/constant/api_type.go b/relay/constant/api_type.go
deleted file mode 100644
index b249f6a2..00000000
--- a/relay/constant/api_type.go
+++ /dev/null
@@ -1,48 +0,0 @@
-package constant
-
-import (
-	"github.com/songquanpeng/one-api/common"
-)
-
-const (
-	APITypeOpenAI = iota
-	APITypeAnthropic
-	APITypePaLM
-	APITypeBaidu
-	APITypeZhipu
-	APITypeAli
-	APITypeXunfei
-	APITypeAIProxyLibrary
-	APITypeTencent
-	APITypeGemini
-	APITypeOllama
-
-	APITypeDummy // this one is only for count, do not add any channel after this
-)
-
-func ChannelType2APIType(channelType int) int {
-	apiType := APITypeOpenAI
-	switch channelType {
-	case common.ChannelTypeAnthropic:
-		apiType = APITypeAnthropic
-	case common.ChannelTypeBaidu:
-		apiType = APITypeBaidu
-	case common.ChannelTypePaLM:
-		apiType = APITypePaLM
-	case common.ChannelTypeZhipu:
-		apiType = APITypeZhipu
-	case common.ChannelTypeAli:
-		apiType = APITypeAli
-	case common.ChannelTypeXunfei:
-		apiType = APITypeXunfei
-	case common.ChannelTypeAIProxyLibrary:
-		apiType = APITypeAIProxyLibrary
-	case common.ChannelTypeTencent:
-		apiType = APITypeTencent
-	case common.ChannelTypeGemini:
-		apiType = APITypeGemini
-	case common.ChannelTypeOllama:
-		apiType = APITypeOllama
-	}
-	return apiType
-}
diff --git a/relay/constant/common.go b/relay/constant/common.go
index b6606cc6..f31477ca 100644
--- a/relay/constant/common.go
+++ b/relay/constant/common.go
@@ -1,3 +1,5 @@
 package constant
 
 var StopFinishReason = "stop"
+var StreamObject = "chat.completion.chunk"
+var NonStreamObject = "chat.completion"
diff --git a/relay/constant/finishreason/define.go b/relay/constant/finishreason/define.go
new file mode 100644
index 00000000..1ed9c425
--- /dev/null
+++ b/relay/constant/finishreason/define.go
@@ -0,0 +1,5 @@
+package finishreason
+
+const (
+	Stop = "stop"
+)
diff --git a/relay/constant/image.go b/relay/constant/image.go
deleted file mode 100644
index 5e04895f..00000000
--- a/relay/constant/image.go
+++ /dev/null
@@ -1,24 +0,0 @@
-package constant
-
-var DalleSizeRatios = map[string]map[string]float64{
-	"dall-e-2": {
-		"256x256":   1,
-		"512x512":   1.125,
-		"1024x1024": 1.25,
-	},
-	"dall-e-3": {
-		"1024x1024": 1,
-		"1024x1792": 2,
-		"1792x1024": 2,
-	},
-}
-
-var DalleGenerationImageAmounts = map[string][2]int{
-	"dall-e-2": {1, 10},
-	"dall-e-3": {1, 1}, // OpenAI allows n=1 currently.
-}
-
-var DalleImagePromptLengthLimitations = map[string]int{
-	"dall-e-2": 1000,
-	"dall-e-3": 4000,
-}
diff --git a/relay/constant/relay_mode.go b/relay/constant/relay_mode.go
deleted file mode 100644
index 5e2fe574..00000000
--- a/relay/constant/relay_mode.go
+++ /dev/null
@@ -1,42 +0,0 @@
-package constant
-
-import "strings"
-
-const (
-	RelayModeUnknown = iota
-	RelayModeChatCompletions
-	RelayModeCompletions
-	RelayModeEmbeddings
-	RelayModeModerations
-	RelayModeImagesGenerations
-	RelayModeEdits
-	RelayModeAudioSpeech
-	RelayModeAudioTranscription
-	RelayModeAudioTranslation
-)
-
-func Path2RelayMode(path string) int {
-	relayMode := RelayModeUnknown
-	if strings.HasPrefix(path, "/v1/chat/completions") {
-		relayMode = RelayModeChatCompletions
-	} else if strings.HasPrefix(path, "/v1/completions") {
-		relayMode = RelayModeCompletions
-	} else if strings.HasPrefix(path, "/v1/embeddings") {
-		relayMode = RelayModeEmbeddings
-	} else if strings.HasSuffix(path, "embeddings") {
-		relayMode = RelayModeEmbeddings
-	} else if strings.HasPrefix(path, "/v1/moderations") {
-		relayMode = RelayModeModerations
-	} else if strings.HasPrefix(path, "/v1/images/generations") {
-		relayMode = RelayModeImagesGenerations
-	} else if strings.HasPrefix(path, "/v1/edits") {
-		relayMode = RelayModeEdits
-	} else if strings.HasPrefix(path, "/v1/audio/speech") {
-		relayMode = RelayModeAudioSpeech
-	} else if strings.HasPrefix(path, "/v1/audio/transcriptions") {
-		relayMode = RelayModeAudioTranscription
-	} else if strings.HasPrefix(path, "/v1/audio/translations") {
-		relayMode = RelayModeAudioTranslation
-	}
-	return relayMode
-}
diff --git a/relay/constant/role/define.go b/relay/constant/role/define.go
new file mode 100644
index 00000000..5097c97e
--- /dev/null
+++ b/relay/constant/role/define.go
@@ -0,0 +1,6 @@
+package role
+
+const (
+	System    = "system"
+	Assistant = "assistant"
+)
diff --git a/relay/controller/audio.go b/relay/controller/audio.go
index e69c47b3..e3d57b1e 100644
--- a/relay/controller/audio.go
+++ b/relay/controller/audio.go
@@ -7,33 +7,40 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
 	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/client"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/billing"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/meta"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 )
 
 func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatusCode {
 	ctx := c.Request.Context()
+	meta := meta.GetByContext(c)
 	audioModel := "whisper-1"
 
-	tokenId := c.GetInt("token_id")
-	channelType := c.GetInt("channel")
-	channelId := c.GetInt("channel_id")
-	userId := c.GetInt("id")
-	group := c.GetString("group")
-	tokenName := c.GetString("token_name")
+	tokenId := c.GetInt(ctxkey.TokenId)
+	channelType := c.GetInt(ctxkey.Channel)
+	channelId := c.GetInt(ctxkey.ChannelId)
+	userId := c.GetInt(ctxkey.Id)
+	group := c.GetString(ctxkey.Group)
+	tokenName := c.GetString(ctxkey.TokenName)
 
 	var ttsRequest openai.TextToSpeechRequest
-	if relayMode == constant.RelayModeAudioSpeech {
+	if relayMode == relaymode.AudioSpeech {
 		// Read JSON
 		err := common.UnmarshalBodyReusable(c, &ttsRequest)
 		// Check if JSON is valid
@@ -47,13 +54,13 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		}
 	}
 
-	modelRatio := common.GetModelRatio(audioModel)
-	groupRatio := common.GetGroupRatio(group)
+	modelRatio := billingratio.GetModelRatio(audioModel, channelType)
+	groupRatio := billingratio.GetGroupRatio(group)
 	ratio := modelRatio * groupRatio
 	var quota int64
 	var preConsumedQuota int64
 	switch relayMode {
-	case constant.RelayModeAudioSpeech:
+	case relaymode.AudioSpeech:
 		preConsumedQuota = int64(float64(len(ttsRequest.Input)) * ratio)
 		quota = preConsumedQuota
 	default:
@@ -83,33 +90,44 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 			return openai.ErrorWrapper(err, "pre_consume_token_quota_failed", http.StatusForbidden)
 		}
 	}
+	succeed := false
+	defer func() {
+		if succeed {
+			return
+		}
+		if preConsumedQuota > 0 {
+			// we need to roll back the pre-consumed quota
+			defer func(ctx context.Context) {
+				go func() {
+					// negative means add quota back for token & user
+					err := model.PostConsumeTokenQuota(tokenId, -preConsumedQuota)
+					if err != nil {
+						logger.Error(ctx, fmt.Sprintf("error rollback pre-consumed quota: %s", err.Error()))
+					}
+				}()
+			}(c.Request.Context())
+		}
+	}()
 
 	// map model name
-	modelMapping := c.GetString("model_mapping")
-	if modelMapping != "" {
-		modelMap := make(map[string]string)
-		err := json.Unmarshal([]byte(modelMapping), &modelMap)
-		if err != nil {
-			return openai.ErrorWrapper(err, "unmarshal_model_mapping_failed", http.StatusInternalServerError)
-		}
-		if modelMap[audioModel] != "" {
-			audioModel = modelMap[audioModel]
-		}
+	modelMapping := c.GetStringMapString(ctxkey.ModelMapping)
+	if modelMapping != nil && modelMapping[audioModel] != "" {
+		audioModel = modelMapping[audioModel]
 	}
 
-	baseURL := common.ChannelBaseURLs[channelType]
+	baseURL := channeltype.ChannelBaseURLs[channelType]
 	requestURL := c.Request.URL.String()
-	if c.GetString("base_url") != "" {
-		baseURL = c.GetString("base_url")
+	if c.GetString(ctxkey.BaseURL) != "" {
+		baseURL = c.GetString(ctxkey.BaseURL)
 	}
 
-	fullRequestURL := util.GetFullRequestURL(baseURL, requestURL, channelType)
-	if channelType == common.ChannelTypeAzure {
-		apiVersion := util.GetAzureAPIVersion(c)
-		if relayMode == constant.RelayModeAudioTranscription {
+	fullRequestURL := openai.GetFullRequestURL(baseURL, requestURL, channelType)
+	if channelType == channeltype.Azure {
+		apiVersion := meta.Config.APIVersion
+		if relayMode == relaymode.AudioTranscription {
 			// https://learn.microsoft.com/en-us/azure/ai-services/openai/whisper-quickstart?tabs=command-line#rest-api
 			fullRequestURL = fmt.Sprintf("%s/openai/deployments/%s/audio/transcriptions?api-version=%s", baseURL, audioModel, apiVersion)
-		} else if relayMode == constant.RelayModeAudioSpeech {
+		} else if relayMode == relaymode.AudioSpeech {
 			// https://learn.microsoft.com/en-us/azure/ai-services/openai/text-to-speech-quickstart?tabs=command-line#rest-api
 			fullRequestURL = fmt.Sprintf("%s/openai/deployments/%s/audio/speech?api-version=%s", baseURL, audioModel, apiVersion)
 		}
@@ -128,7 +146,7 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		return openai.ErrorWrapper(err, "new_request_failed", http.StatusInternalServerError)
 	}
 
-	if (relayMode == constant.RelayModeAudioTranscription || relayMode == constant.RelayModeAudioSpeech) && channelType == common.ChannelTypeAzure {
+	if (relayMode == relaymode.AudioTranscription || relayMode == relaymode.AudioSpeech) && channelType == channeltype.Azure {
 		// https://learn.microsoft.com/en-us/azure/ai-services/openai/whisper-quickstart?tabs=command-line#rest-api
 		apiKey := c.Request.Header.Get("Authorization")
 		apiKey = strings.TrimPrefix(apiKey, "Bearer ")
@@ -140,7 +158,7 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
 	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
 
-	resp, err := util.HTTPClient.Do(req)
+	resp, err := client.HTTPClient.Do(req)
 	if err != nil {
 		return openai.ErrorWrapper(err, "do_request_failed", http.StatusInternalServerError)
 	}
@@ -154,7 +172,7 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		return openai.ErrorWrapper(err, "close_request_body_failed", http.StatusInternalServerError)
 	}
 
-	if relayMode != constant.RelayModeAudioSpeech {
+	if relayMode != relaymode.AudioSpeech {
 		responseBody, err := io.ReadAll(resp.Body)
 		if err != nil {
 			return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
@@ -193,23 +211,12 @@ func RelayAudioHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		resp.Body = io.NopCloser(bytes.NewBuffer(responseBody))
 	}
 	if resp.StatusCode != http.StatusOK {
-		if preConsumedQuota > 0 {
-			// we need to roll back the pre-consumed quota
-			defer func(ctx context.Context) {
-				go func() {
-					// negative means add quota back for token & user
-					err := model.PostConsumeTokenQuota(tokenId, -preConsumedQuota)
-					if err != nil {
-						logger.Error(ctx, fmt.Sprintf("error rollback pre-consumed quota: %s", err.Error()))
-					}
-				}()
-			}(c.Request.Context())
-		}
-		return util.RelayErrorHandler(resp)
+		return RelayErrorHandler(resp)
 	}
+	succeed = true
 	quotaDelta := quota - preConsumedQuota
 	defer func(ctx context.Context) {
-		go util.PostConsumeQuota(ctx, tokenId, quotaDelta, quota, userId, channelId, modelRatio, groupRatio, audioModel, tokenName)
+		go billing.PostConsumeQuota(ctx, tokenId, quotaDelta, quota, userId, channelId, modelRatio, groupRatio, audioModel, tokenName)
 	}(c.Request.Context())
 
 	for k, v := range resp.Header {
diff --git a/relay/controller/error.go b/relay/controller/error.go
new file mode 100644
index 00000000..29d4f125
--- /dev/null
+++ b/relay/controller/error.go
@@ -0,0 +1,101 @@
+package controller
+
+import (
+	"encoding/json"
+	"fmt"
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay/model"
+	"io"
+	"net/http"
+	"strconv"
+)
+
+type GeneralErrorResponse struct {
+	Error    model.Error `json:"error"`
+	Message  string      `json:"message"`
+	Msg      string      `json:"msg"`
+	Err      string      `json:"err"`
+	ErrorMsg string      `json:"error_msg"`
+	Header   struct {
+		Message string `json:"message"`
+	} `json:"header"`
+	Response struct {
+		Error struct {
+			Message string `json:"message"`
+		} `json:"error"`
+	} `json:"response"`
+}
+
+func (e GeneralErrorResponse) ToMessage() string {
+	if e.Error.Message != "" {
+		return e.Error.Message
+	}
+	if e.Message != "" {
+		return e.Message
+	}
+	if e.Msg != "" {
+		return e.Msg
+	}
+	if e.Err != "" {
+		return e.Err
+	}
+	if e.ErrorMsg != "" {
+		return e.ErrorMsg
+	}
+	if e.Header.Message != "" {
+		return e.Header.Message
+	}
+	if e.Response.Error.Message != "" {
+		return e.Response.Error.Message
+	}
+	return ""
+}
+
+func RelayErrorHandler(resp *http.Response) (ErrorWithStatusCode *model.ErrorWithStatusCode) {
+	if resp == nil {
+		return &model.ErrorWithStatusCode{
+			StatusCode: 500,
+			Error: model.Error{
+				Message: "resp is nil",
+				Type:    "upstream_error",
+				Code:    "bad_response",
+			},
+		}
+	}
+	ErrorWithStatusCode = &model.ErrorWithStatusCode{
+		StatusCode: resp.StatusCode,
+		Error: model.Error{
+			Message: "",
+			Type:    "upstream_error",
+			Code:    "bad_response_status_code",
+			Param:   strconv.Itoa(resp.StatusCode),
+		},
+	}
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return
+	}
+	if config.DebugEnabled {
+		logger.SysLog(fmt.Sprintf("error happened, status code: %d, response: \n%s", resp.StatusCode, string(responseBody)))
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return
+	}
+	var errResponse GeneralErrorResponse
+	err = json.Unmarshal(responseBody, &errResponse)
+	if err != nil {
+		return
+	}
+	if errResponse.Error.Message != "" {
+		// OpenAI format error, so we override the default one
+		ErrorWithStatusCode.Error = errResponse.Error
+	} else {
+		ErrorWithStatusCode.Error.Message = errResponse.ToMessage()
+	}
+	if ErrorWithStatusCode.Error.Message == "" {
+		ErrorWithStatusCode.Error.Message = fmt.Sprintf("bad response status code %d", resp.StatusCode)
+	}
+	return
+}
diff --git a/relay/controller/helper.go b/relay/controller/helper.go
index 600a8d65..5f5fc90c 100644
--- a/relay/controller/helper.go
+++ b/relay/controller/helper.go
@@ -4,17 +4,23 @@ import (
 	"context"
 	"errors"
 	"fmt"
+	"github.com/songquanpeng/one-api/relay/constant/role"
+	"math"
+	"net/http"
+	"strings"
+
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/controller/validator"
+	"github.com/songquanpeng/one-api/relay/meta"
 	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
-	"math"
-	"net/http"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 )
 
 func getAndValidateTextRequest(c *gin.Context, relayMode int) (*relaymodel.GeneralOpenAIRequest, error) {
@@ -23,99 +29,40 @@ func getAndValidateTextRequest(c *gin.Context, relayMode int) (*relaymodel.Gener
 	if err != nil {
 		return nil, err
 	}
-	if relayMode == constant.RelayModeModerations && textRequest.Model == "" {
+	if relayMode == relaymode.Moderations && textRequest.Model == "" {
 		textRequest.Model = "text-moderation-latest"
 	}
-	if relayMode == constant.RelayModeEmbeddings && textRequest.Model == "" {
+	if relayMode == relaymode.Embeddings && textRequest.Model == "" {
 		textRequest.Model = c.Param("model")
 	}
-	err = util.ValidateTextRequest(textRequest, relayMode)
+	err = validator.ValidateTextRequest(textRequest, relayMode)
 	if err != nil {
 		return nil, err
 	}
 	return textRequest, nil
 }
 
-func getImageRequest(c *gin.Context, relayMode int) (*openai.ImageRequest, error) {
-	imageRequest := &openai.ImageRequest{}
-	err := common.UnmarshalBodyReusable(c, imageRequest)
-	if err != nil {
-		return nil, err
-	}
-	if imageRequest.N == 0 {
-		imageRequest.N = 1
-	}
-	if imageRequest.Size == "" {
-		imageRequest.Size = "1024x1024"
-	}
-	if imageRequest.Model == "" {
-		imageRequest.Model = "dall-e-2"
-	}
-	return imageRequest, nil
-}
-
-func validateImageRequest(imageRequest *openai.ImageRequest, meta *util.RelayMeta) *relaymodel.ErrorWithStatusCode {
-	// model validation
-	_, hasValidSize := constant.DalleSizeRatios[imageRequest.Model][imageRequest.Size]
-	if !hasValidSize {
-		return openai.ErrorWrapper(errors.New("size not supported for this image model"), "size_not_supported", http.StatusBadRequest)
-	}
-	// check prompt length
-	if imageRequest.Prompt == "" {
-		return openai.ErrorWrapper(errors.New("prompt is required"), "prompt_missing", http.StatusBadRequest)
-	}
-	if len(imageRequest.Prompt) > constant.DalleImagePromptLengthLimitations[imageRequest.Model] {
-		return openai.ErrorWrapper(errors.New("prompt is too long"), "prompt_too_long", http.StatusBadRequest)
-	}
-	// Number of generated images validation
-	if !isWithinRange(imageRequest.Model, imageRequest.N) {
-		// channel not azure
-		if meta.ChannelType != common.ChannelTypeAzure {
-			return openai.ErrorWrapper(errors.New("invalid value of n"), "n_not_within_range", http.StatusBadRequest)
-		}
-	}
-	return nil
-}
-
-func getImageCostRatio(imageRequest *openai.ImageRequest) (float64, error) {
-	if imageRequest == nil {
-		return 0, errors.New("imageRequest is nil")
-	}
-	imageCostRatio, hasValidSize := constant.DalleSizeRatios[imageRequest.Model][imageRequest.Size]
-	if !hasValidSize {
-		return 0, fmt.Errorf("size not supported for this image model: %s", imageRequest.Size)
-	}
-	if imageRequest.Quality == "hd" && imageRequest.Model == "dall-e-3" {
-		if imageRequest.Size == "1024x1024" {
-			imageCostRatio *= 2
-		} else {
-			imageCostRatio *= 1.5
-		}
-	}
-	return imageCostRatio, nil
-}
-
 func getPromptTokens(textRequest *relaymodel.GeneralOpenAIRequest, relayMode int) int {
 	switch relayMode {
-	case constant.RelayModeChatCompletions:
+	case relaymode.ChatCompletions:
 		return openai.CountTokenMessages(textRequest.Messages, textRequest.Model)
-	case constant.RelayModeCompletions:
+	case relaymode.Completions:
 		return openai.CountTokenInput(textRequest.Prompt, textRequest.Model)
-	case constant.RelayModeModerations:
+	case relaymode.Moderations:
 		return openai.CountTokenInput(textRequest.Input, textRequest.Model)
 	}
 	return 0
 }
 
 func getPreConsumedQuota(textRequest *relaymodel.GeneralOpenAIRequest, promptTokens int, ratio float64) int64 {
-	preConsumedTokens := config.PreConsumedQuota
+	preConsumedTokens := config.PreConsumedQuota + int64(promptTokens)
 	if textRequest.MaxTokens != 0 {
-		preConsumedTokens = int64(promptTokens) + int64(textRequest.MaxTokens)
+		preConsumedTokens += int64(textRequest.MaxTokens)
 	}
 	return int64(float64(preConsumedTokens) * ratio)
 }
 
-func preConsumeQuota(ctx context.Context, textRequest *relaymodel.GeneralOpenAIRequest, promptTokens int, ratio float64, meta *util.RelayMeta) (int64, *relaymodel.ErrorWithStatusCode) {
+func preConsumeQuota(ctx context.Context, textRequest *relaymodel.GeneralOpenAIRequest, promptTokens int, ratio float64, meta *meta.Meta) (int64, *relaymodel.ErrorWithStatusCode) {
 	preConsumedQuota := getPreConsumedQuota(textRequest, promptTokens, ratio)
 
 	userQuota, err := model.CacheGetUserQuota(ctx, meta.UserId)
@@ -144,13 +91,13 @@ func preConsumeQuota(ctx context.Context, textRequest *relaymodel.GeneralOpenAIR
 	return preConsumedQuota, nil
 }
 
-func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *util.RelayMeta, textRequest *relaymodel.GeneralOpenAIRequest, ratio float64, preConsumedQuota int64, modelRatio float64, groupRatio float64) {
+func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *meta.Meta, textRequest *relaymodel.GeneralOpenAIRequest, ratio float64, preConsumedQuota int64, modelRatio float64, groupRatio float64, systemPromptReset bool) {
 	if usage == nil {
 		logger.Error(ctx, "usage is nil, which is unexpected")
 		return
 	}
 	var quota int64
-	completionRatio := common.GetCompletionRatio(textRequest.Model)
+	completionRatio := billingratio.GetCompletionRatio(textRequest.Model, meta.ChannelType)
 	promptTokens := usage.PromptTokens
 	completionTokens := usage.CompletionTokens
 	quota = int64(math.Ceil((float64(promptTokens) + float64(completionTokens)*completionRatio) * ratio))
@@ -172,8 +119,69 @@ func postConsumeQuota(ctx context.Context, usage *relaymodel.Usage, meta *util.R
 	if err != nil {
 		logger.Error(ctx, "error update user quota cache: "+err.Error())
 	}
-	logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f，补全倍率 %.2f", modelRatio, groupRatio, completionRatio)
+	var extraLog string
+	if systemPromptReset {
+		extraLog = " （注意系统提示词已被重置）"
+	}
+	logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f，补全倍率 %.2f%s", modelRatio, groupRatio, completionRatio, extraLog)
 	model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, promptTokens, completionTokens, textRequest.Model, meta.TokenName, quota, logContent)
 	model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
 	model.UpdateChannelUsedQuota(meta.ChannelId, quota)
 }
+
+func getMappedModelName(modelName string, mapping map[string]string) (string, bool) {
+	if mapping == nil {
+		return modelName, false
+	}
+	mappedModelName := mapping[modelName]
+	if mappedModelName != "" {
+		return mappedModelName, true
+	}
+	return modelName, false
+}
+
+func isErrorHappened(meta *meta.Meta, resp *http.Response) bool {
+	if resp == nil {
+		if meta.ChannelType == channeltype.AwsClaude {
+			return false
+		}
+		return true
+	}
+	if resp.StatusCode != http.StatusOK &&
+		// replicate return 201 to create a task
+		resp.StatusCode != http.StatusCreated {
+		return true
+	}
+	if meta.ChannelType == channeltype.DeepL {
+		// skip stream check for deepl
+		return false
+	}
+
+	if meta.IsStream && strings.HasPrefix(resp.Header.Get("Content-Type"), "application/json") &&
+		// Even if stream mode is enabled, replicate will first return a task info in JSON format,
+		// requiring the client to request the stream endpoint in the task info
+		meta.ChannelType != channeltype.Replicate {
+		return true
+	}
+	return false
+}
+
+func setSystemPrompt(ctx context.Context, request *relaymodel.GeneralOpenAIRequest, prompt string) (reset bool) {
+	if prompt == "" {
+		return false
+	}
+	if len(request.Messages) == 0 {
+		return false
+	}
+	if request.Messages[0].Role == role.System {
+		request.Messages[0].Content = prompt
+		logger.Infof(ctx, "rewrite system prompt")
+		return true
+	}
+	request.Messages = append([]relaymodel.Message{{
+		Role:    role.System,
+		Content: prompt,
+	}}, request.Messages...)
+	logger.Infof(ctx, "add system prompt")
+	return true
+}
diff --git a/relay/controller/image.go b/relay/controller/image.go
index d81dadf6..1b69d97d 100644
--- a/relay/controller/image.go
+++ b/relay/controller/image.go
@@ -6,33 +6,105 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/model"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
 	"io"
 	"net/http"
-	"strings"
 
 	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/meta"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
 )
 
-func isWithinRange(element string, value int) bool {
-	if _, ok := constant.DalleGenerationImageAmounts[element]; !ok {
-		return false
+func getImageRequest(c *gin.Context, _ int) (*relaymodel.ImageRequest, error) {
+	imageRequest := &relaymodel.ImageRequest{}
+	err := common.UnmarshalBodyReusable(c, imageRequest)
+	if err != nil {
+		return nil, err
 	}
-	min := constant.DalleGenerationImageAmounts[element][0]
-	max := constant.DalleGenerationImageAmounts[element][1]
+	if imageRequest.N == 0 {
+		imageRequest.N = 1
+	}
+	if imageRequest.Size == "" {
+		imageRequest.Size = "1024x1024"
+	}
+	if imageRequest.Model == "" {
+		imageRequest.Model = "dall-e-2"
+	}
+	return imageRequest, nil
+}
 
-	return value >= min && value <= max
+func isValidImageSize(model string, size string) bool {
+	if model == "cogview-3" || billingratio.ImageSizeRatios[model] == nil {
+		return true
+	}
+	_, ok := billingratio.ImageSizeRatios[model][size]
+	return ok
+}
+
+func isValidImagePromptLength(model string, promptLength int) bool {
+	maxPromptLength, ok := billingratio.ImagePromptLengthLimitations[model]
+	return !ok || promptLength <= maxPromptLength
+}
+
+func isWithinRange(element string, value int) bool {
+	amounts, ok := billingratio.ImageGenerationAmounts[element]
+	return !ok || (value >= amounts[0] && value <= amounts[1])
+}
+
+func getImageSizeRatio(model string, size string) float64 {
+	if ratio, ok := billingratio.ImageSizeRatios[model][size]; ok {
+		return ratio
+	}
+	return 1
+}
+
+func validateImageRequest(imageRequest *relaymodel.ImageRequest, _ *meta.Meta) *relaymodel.ErrorWithStatusCode {
+	// check prompt length
+	if imageRequest.Prompt == "" {
+		return openai.ErrorWrapper(errors.New("prompt is required"), "prompt_missing", http.StatusBadRequest)
+	}
+
+	// model validation
+	if !isValidImageSize(imageRequest.Model, imageRequest.Size) {
+		return openai.ErrorWrapper(errors.New("size not supported for this image model"), "size_not_supported", http.StatusBadRequest)
+	}
+
+	if !isValidImagePromptLength(imageRequest.Model, len(imageRequest.Prompt)) {
+		return openai.ErrorWrapper(errors.New("prompt is too long"), "prompt_too_long", http.StatusBadRequest)
+	}
+
+	// Number of generated images validation
+	if !isWithinRange(imageRequest.Model, imageRequest.N) {
+		return openai.ErrorWrapper(errors.New("invalid value of n"), "n_not_within_range", http.StatusBadRequest)
+	}
+	return nil
+}
+
+func getImageCostRatio(imageRequest *relaymodel.ImageRequest) (float64, error) {
+	if imageRequest == nil {
+		return 0, errors.New("imageRequest is nil")
+	}
+	imageCostRatio := getImageSizeRatio(imageRequest.Model, imageRequest.Size)
+	if imageRequest.Quality == "hd" && imageRequest.Model == "dall-e-3" {
+		if imageRequest.Size == "1024x1024" {
+			imageCostRatio *= 2
+		} else {
+			imageCostRatio *= 1.5
+		}
+	}
+	return imageCostRatio, nil
 }
 
 func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatusCode {
 	ctx := c.Request.Context()
-	meta := util.GetRelayMeta(c)
+	meta := meta.GetByContext(c)
 	imageRequest, err := getImageRequest(c, meta.Mode)
 	if err != nil {
 		logger.Errorf(ctx, "getImageRequest failed: %s", err.Error())
@@ -42,7 +114,7 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 	// map model name
 	var isModelMapped bool
 	meta.OriginModelName = imageRequest.Model
-	imageRequest.Model, isModelMapped = util.GetMappedModelName(imageRequest.Model, meta.ModelMapping)
+	imageRequest.Model, isModelMapped = getMappedModelName(imageRequest.Model, meta.ModelMapping)
 	meta.ActualModelName = imageRequest.Model
 
 	// model validation
@@ -56,17 +128,13 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		return openai.ErrorWrapper(err, "get_image_cost_ratio_failed", http.StatusInternalServerError)
 	}
 
-	requestURL := c.Request.URL.String()
-	fullRequestURL := util.GetFullRequestURL(meta.BaseURL, requestURL, meta.ChannelType)
-	if meta.ChannelType == common.ChannelTypeAzure {
-		// https://learn.microsoft.com/en-us/azure/ai-services/openai/dall-e-quickstart?tabs=dalle3%2Ccommand-line&pivots=rest-api
-		apiVersion := util.GetAzureAPIVersion(c)
-		// https://{resource_name}.openai.azure.com/openai/deployments/dall-e-3/images/generations?api-version=2024-03-01-preview
-		fullRequestURL = fmt.Sprintf("%s/openai/deployments/%s/images/generations?api-version=%s", meta.BaseURL, imageRequest.Model, apiVersion)
-	}
+	imageModel := imageRequest.Model
+	// Convert the original image model
+	imageRequest.Model, _ = getMappedModelName(imageRequest.Model, billingratio.ImageOriginModelName)
+	c.Set("response_format", imageRequest.ResponseFormat)
 
 	var requestBody io.Reader
-	if isModelMapped || meta.ChannelType == common.ChannelTypeAzure { // make Azure channel request body
+	if isModelMapped || meta.ChannelType == channeltype.Azure { // make Azure channel request body
 		jsonStr, err := json.Marshal(imageRequest)
 		if err != nil {
 			return openai.ErrorWrapper(err, "marshal_image_request_failed", http.StatusInternalServerError)
@@ -76,51 +144,61 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 		requestBody = c.Request.Body
 	}
 
-	modelRatio := common.GetModelRatio(imageRequest.Model)
-	groupRatio := common.GetGroupRatio(meta.Group)
+	adaptor := relay.GetAdaptor(meta.APIType)
+	if adaptor == nil {
+		return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
+	}
+	adaptor.Init(meta)
+
+	// these adaptors need to convert the request
+	switch meta.ChannelType {
+	case channeltype.Zhipu,
+		channeltype.Ali,
+		channeltype.Replicate,
+		channeltype.Baidu:
+		finalRequest, err := adaptor.ConvertImageRequest(imageRequest)
+		if err != nil {
+			return openai.ErrorWrapper(err, "convert_image_request_failed", http.StatusInternalServerError)
+		}
+		jsonStr, err := json.Marshal(finalRequest)
+		if err != nil {
+			return openai.ErrorWrapper(err, "marshal_image_request_failed", http.StatusInternalServerError)
+		}
+		requestBody = bytes.NewBuffer(jsonStr)
+	}
+
+	modelRatio := billingratio.GetModelRatio(imageModel, meta.ChannelType)
+	groupRatio := billingratio.GetGroupRatio(meta.Group)
 	ratio := modelRatio * groupRatio
 	userQuota, err := model.CacheGetUserQuota(ctx, meta.UserId)
 
-	quota := int64(ratio*imageCostRatio*1000) * int64(imageRequest.N)
+	var quota int64
+	switch meta.ChannelType {
+	case channeltype.Replicate:
+		// replicate always return 1 image
+		quota = int64(ratio * imageCostRatio * 1000)
+	default:
+		quota = int64(ratio*imageCostRatio*1000) * int64(imageRequest.N)
+	}
 
 	if userQuota-quota < 0 {
 		return openai.ErrorWrapper(errors.New("user quota is not enough"), "insufficient_user_quota", http.StatusForbidden)
 	}
 
-	req, err := http.NewRequest(c.Request.Method, fullRequestURL, requestBody)
-	if err != nil {
-		return openai.ErrorWrapper(err, "new_request_failed", http.StatusInternalServerError)
-	}
-	token := c.Request.Header.Get("Authorization")
-	if meta.ChannelType == common.ChannelTypeAzure { // Azure authentication
-		token = strings.TrimPrefix(token, "Bearer ")
-		req.Header.Set("api-key", token)
-	} else {
-		req.Header.Set("Authorization", token)
-	}
-
-	req.Header.Set("Content-Type", c.Request.Header.Get("Content-Type"))
-	req.Header.Set("Accept", c.Request.Header.Get("Accept"))
-
-	resp, err := util.HTTPClient.Do(req)
+	// do request
+	resp, err := adaptor.DoRequest(c, meta, requestBody)
 	if err != nil {
+		logger.Errorf(ctx, "DoRequest failed: %s", err.Error())
 		return openai.ErrorWrapper(err, "do_request_failed", http.StatusInternalServerError)
 	}
 
-	err = req.Body.Close()
-	if err != nil {
-		return openai.ErrorWrapper(err, "close_request_body_failed", http.StatusInternalServerError)
-	}
-	err = c.Request.Body.Close()
-	if err != nil {
-		return openai.ErrorWrapper(err, "close_request_body_failed", http.StatusInternalServerError)
-	}
-	var imageResponse openai.ImageResponse
-
 	defer func(ctx context.Context) {
-		if resp.StatusCode != http.StatusOK {
+		if resp != nil &&
+			resp.StatusCode != http.StatusCreated && // replicate returns 201
+			resp.StatusCode != http.StatusOK {
 			return
 		}
+
 		err := model.PostConsumeTokenQuota(meta.TokenId, quota)
 		if err != nil {
 			logger.SysError("error consuming token remain quota: " + err.Error())
@@ -130,43 +208,21 @@ func RelayImageHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatus
 			logger.SysError("error update user quota cache: " + err.Error())
 		}
 		if quota != 0 {
-			tokenName := c.GetString("token_name")
+			tokenName := c.GetString(ctxkey.TokenName)
 			logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
 			model.RecordConsumeLog(ctx, meta.UserId, meta.ChannelId, 0, 0, imageRequest.Model, tokenName, quota, logContent)
 			model.UpdateUserUsedQuotaAndRequestCount(meta.UserId, quota)
-			channelId := c.GetInt("channel_id")
+			channelId := c.GetInt(ctxkey.ChannelId)
 			model.UpdateChannelUsedQuota(channelId, quota)
 		}
 	}(c.Request.Context())
 
-	responseBody, err := io.ReadAll(resp.Body)
-
-	if err != nil {
-		return openai.ErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
-	}
-	err = resp.Body.Close()
-	if err != nil {
-		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
-	}
-	err = json.Unmarshal(responseBody, &imageResponse)
-	if err != nil {
-		return openai.ErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError)
+	// do response
+	_, respErr := adaptor.DoResponse(c, resp, meta)
+	if respErr != nil {
+		logger.Errorf(ctx, "respErr is not nil: %+v", respErr)
+		return respErr
 	}
 
-	resp.Body = io.NopCloser(bytes.NewBuffer(responseBody))
-
-	for k, v := range resp.Header {
-		c.Writer.Header().Set(k, v[0])
-	}
-	c.Writer.WriteHeader(resp.StatusCode)
-
-	_, err = io.Copy(c.Writer, resp.Body)
-	if err != nil {
-		return openai.ErrorWrapper(err, "copy_response_body_failed", http.StatusInternalServerError)
-	}
-	err = resp.Body.Close()
-	if err != nil {
-		return openai.ErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
-	}
 	return nil
 }
diff --git a/relay/controller/proxy.go b/relay/controller/proxy.go
new file mode 100644
index 00000000..dcaf15a9
--- /dev/null
+++ b/relay/controller/proxy.go
@@ -0,0 +1,41 @@
+// Package controller is a package for handling the relay controller
+package controller
+
+import (
+	"fmt"
+	"net/http"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/meta"
+	relaymodel "github.com/songquanpeng/one-api/relay/model"
+)
+
+// RelayProxyHelper is a helper function to proxy the request to the upstream service
+func RelayProxyHelper(c *gin.Context, relayMode int) *relaymodel.ErrorWithStatusCode {
+	ctx := c.Request.Context()
+	meta := meta.GetByContext(c)
+
+	adaptor := relay.GetAdaptor(meta.APIType)
+	if adaptor == nil {
+		return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
+	}
+	adaptor.Init(meta)
+
+	resp, err := adaptor.DoRequest(c, meta, c.Request.Body)
+	if err != nil {
+		logger.Errorf(ctx, "DoRequest failed: %s", err.Error())
+		return openai.ErrorWrapper(err, "do_request_failed", http.StatusInternalServerError)
+	}
+
+	// do response
+	_, respErr := adaptor.DoResponse(c, resp, meta)
+	if respErr != nil {
+		logger.Errorf(ctx, "respErr is not nil: %+v", respErr)
+		return respErr
+	}
+
+	return nil
+}
diff --git a/relay/controller/text.go b/relay/controller/text.go
index ba008713..9a47c58b 100644
--- a/relay/controller/text.go
+++ b/relay/controller/text.go
@@ -4,22 +4,26 @@ import (
 	"bytes"
 	"encoding/json"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"github.com/songquanpeng/one-api/relay/helper"
-	"github.com/songquanpeng/one-api/relay/model"
-	"github.com/songquanpeng/one-api/relay/util"
+	"github.com/songquanpeng/one-api/common/config"
 	"io"
 	"net/http"
-	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/logger"
+	"github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/adaptor"
+	"github.com/songquanpeng/one-api/relay/adaptor/openai"
+	"github.com/songquanpeng/one-api/relay/apitype"
+	"github.com/songquanpeng/one-api/relay/billing"
+	billingratio "github.com/songquanpeng/one-api/relay/billing/ratio"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/meta"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	ctx := c.Request.Context()
-	meta := util.GetRelayMeta(c)
+	meta := meta.GetByContext(c)
 	// get & validate textRequest
 	textRequest, err := getAndValidateTextRequest(c, meta.Mode)
 	if err != nil {
@@ -29,13 +33,14 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	meta.IsStream = textRequest.Stream
 
 	// map model name
-	var isModelMapped bool
 	meta.OriginModelName = textRequest.Model
-	textRequest.Model, isModelMapped = util.GetMappedModelName(textRequest.Model, meta.ModelMapping)
+	textRequest.Model, _ = getMappedModelName(textRequest.Model, meta.ModelMapping)
 	meta.ActualModelName = textRequest.Model
+	// set system prompt if not empty
+	systemPromptReset := setSystemPrompt(ctx, textRequest, meta.SystemPrompt)
 	// get model ratio & group ratio
-	modelRatio := common.GetModelRatio(textRequest.Model)
-	groupRatio := common.GetGroupRatio(meta.Group)
+	modelRatio := billingratio.GetModelRatio(textRequest.Model, meta.ChannelType)
+	groupRatio := billingratio.GetGroupRatio(meta.Group)
 	ratio := modelRatio * groupRatio
 	// pre-consume quota
 	promptTokens := getPromptTokens(textRequest, meta.Mode)
@@ -46,36 +51,16 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 		return bizErr
 	}
 
-	adaptor := helper.GetAdaptor(meta.APIType)
+	adaptor := relay.GetAdaptor(meta.APIType)
 	if adaptor == nil {
 		return openai.ErrorWrapper(fmt.Errorf("invalid api type: %d", meta.APIType), "invalid_api_type", http.StatusBadRequest)
 	}
+	adaptor.Init(meta)
 
 	// get request body
-	var requestBody io.Reader
-	if meta.APIType == constant.APITypeOpenAI {
-		// no need to convert request for openai
-		shouldResetRequestBody := isModelMapped || meta.ChannelType == common.ChannelTypeBaichuan // frequency_penalty 0 is not acceptable for baichuan
-		if shouldResetRequestBody {
-			jsonStr, err := json.Marshal(textRequest)
-			if err != nil {
-				return openai.ErrorWrapper(err, "json_marshal_failed", http.StatusInternalServerError)
-			}
-			requestBody = bytes.NewBuffer(jsonStr)
-		} else {
-			requestBody = c.Request.Body
-		}
-	} else {
-		convertedRequest, err := adaptor.ConvertRequest(c, meta.Mode, textRequest)
-		if err != nil {
-			return openai.ErrorWrapper(err, "convert_request_failed", http.StatusInternalServerError)
-		}
-		jsonData, err := json.Marshal(convertedRequest)
-		if err != nil {
-			return openai.ErrorWrapper(err, "json_marshal_failed", http.StatusInternalServerError)
-		}
-		logger.Debugf(ctx, "converted request: \n%s", string(jsonData))
-		requestBody = bytes.NewBuffer(jsonData)
+	requestBody, err := getRequestBody(c, meta, textRequest, adaptor)
+	if err != nil {
+		return openai.ErrorWrapper(err, "convert_request_failed", http.StatusInternalServerError)
 	}
 
 	// do request
@@ -84,21 +69,42 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 		logger.Errorf(ctx, "DoRequest failed: %s", err.Error())
 		return openai.ErrorWrapper(err, "do_request_failed", http.StatusInternalServerError)
 	}
-	errorHappened := (resp.StatusCode != http.StatusOK) || (meta.IsStream && resp.Header.Get("Content-Type") == "application/json")
-	if errorHappened {
-		util.ReturnPreConsumedQuota(ctx, preConsumedQuota, meta.TokenId)
-		return util.RelayErrorHandler(resp)
+	if isErrorHappened(meta, resp) {
+		billing.ReturnPreConsumedQuota(ctx, preConsumedQuota, meta.TokenId)
+		return RelayErrorHandler(resp)
 	}
-	meta.IsStream = meta.IsStream || strings.HasPrefix(resp.Header.Get("Content-Type"), "text/event-stream")
 
 	// do response
 	usage, respErr := adaptor.DoResponse(c, resp, meta)
 	if respErr != nil {
 		logger.Errorf(ctx, "respErr is not nil: %+v", respErr)
-		util.ReturnPreConsumedQuota(ctx, preConsumedQuota, meta.TokenId)
+		billing.ReturnPreConsumedQuota(ctx, preConsumedQuota, meta.TokenId)
 		return respErr
 	}
 	// post-consume quota
-	go postConsumeQuota(ctx, usage, meta, textRequest, ratio, preConsumedQuota, modelRatio, groupRatio)
+	go postConsumeQuota(ctx, usage, meta, textRequest, ratio, preConsumedQuota, modelRatio, groupRatio, systemPromptReset)
 	return nil
 }
+
+func getRequestBody(c *gin.Context, meta *meta.Meta, textRequest *model.GeneralOpenAIRequest, adaptor adaptor.Adaptor) (io.Reader, error) {
+	if !config.EnforceIncludeUsage && meta.APIType == apitype.OpenAI && meta.OriginModelName == meta.ActualModelName && meta.ChannelType != channeltype.Baichuan {
+		// no need to convert request for openai
+		return c.Request.Body, nil
+	}
+
+	// get request body
+	var requestBody io.Reader
+	convertedRequest, err := adaptor.ConvertRequest(c, meta.Mode, textRequest)
+	if err != nil {
+		logger.Debugf(c.Request.Context(), "converted request failed: %s\n", err.Error())
+		return nil, err
+	}
+	jsonData, err := json.Marshal(convertedRequest)
+	if err != nil {
+		logger.Debugf(c.Request.Context(), "converted request json_marshal_failed: %s\n", err.Error())
+		return nil, err
+	}
+	logger.Debugf(c.Request.Context(), "converted request: \n%s", string(jsonData))
+	requestBody = bytes.NewBuffer(jsonData)
+	return requestBody, nil
+}
diff --git a/relay/util/validation.go b/relay/controller/validator/validation.go
similarity index 76%
rename from relay/util/validation.go
rename to relay/controller/validator/validation.go
index ef8d840c..8ff520b8 100644
--- a/relay/util/validation.go
+++ b/relay/controller/validator/validation.go
@@ -1,9 +1,9 @@
-package util
+package validator
 
 import (
 	"errors"
-	"github.com/songquanpeng/one-api/relay/constant"
 	"github.com/songquanpeng/one-api/relay/model"
+	"github.com/songquanpeng/one-api/relay/relaymode"
 	"math"
 )
 
@@ -15,20 +15,20 @@ func ValidateTextRequest(textRequest *model.GeneralOpenAIRequest, relayMode int)
 		return errors.New("model is required")
 	}
 	switch relayMode {
-	case constant.RelayModeCompletions:
+	case relaymode.Completions:
 		if textRequest.Prompt == "" {
 			return errors.New("field prompt is required")
 		}
-	case constant.RelayModeChatCompletions:
+	case relaymode.ChatCompletions:
 		if textRequest.Messages == nil || len(textRequest.Messages) == 0 {
 			return errors.New("field messages is required")
 		}
-	case constant.RelayModeEmbeddings:
-	case constant.RelayModeModerations:
+	case relaymode.Embeddings:
+	case relaymode.Moderations:
 		if textRequest.Input == "" {
 			return errors.New("field input is required")
 		}
-	case constant.RelayModeEdits:
+	case relaymode.Edits:
 		if textRequest.Instruction == "" {
 			return errors.New("field instruction is required")
 		}
diff --git a/relay/helper/main.go b/relay/helper/main.go
deleted file mode 100644
index e7342329..00000000
--- a/relay/helper/main.go
+++ /dev/null
@@ -1,45 +0,0 @@
-package helper
-
-import (
-	"github.com/songquanpeng/one-api/relay/channel"
-	"github.com/songquanpeng/one-api/relay/channel/aiproxy"
-	"github.com/songquanpeng/one-api/relay/channel/ali"
-	"github.com/songquanpeng/one-api/relay/channel/anthropic"
-	"github.com/songquanpeng/one-api/relay/channel/baidu"
-	"github.com/songquanpeng/one-api/relay/channel/gemini"
-	"github.com/songquanpeng/one-api/relay/channel/ollama"
-	"github.com/songquanpeng/one-api/relay/channel/openai"
-	"github.com/songquanpeng/one-api/relay/channel/palm"
-	"github.com/songquanpeng/one-api/relay/channel/tencent"
-	"github.com/songquanpeng/one-api/relay/channel/xunfei"
-	"github.com/songquanpeng/one-api/relay/channel/zhipu"
-	"github.com/songquanpeng/one-api/relay/constant"
-)
-
-func GetAdaptor(apiType int) channel.Adaptor {
-	switch apiType {
-	case constant.APITypeAIProxyLibrary:
-		return &aiproxy.Adaptor{}
-	case constant.APITypeAli:
-		return &ali.Adaptor{}
-	case constant.APITypeAnthropic:
-		return &anthropic.Adaptor{}
-	case constant.APITypeBaidu:
-		return &baidu.Adaptor{}
-	case constant.APITypeGemini:
-		return &gemini.Adaptor{}
-	case constant.APITypeOpenAI:
-		return &openai.Adaptor{}
-	case constant.APITypePaLM:
-		return &palm.Adaptor{}
-	case constant.APITypeTencent:
-		return &tencent.Adaptor{}
-	case constant.APITypeXunfei:
-		return &xunfei.Adaptor{}
-	case constant.APITypeZhipu:
-		return &zhipu.Adaptor{}
-	case constant.APITypeOllama:
-		return &ollama.Adaptor{}
-	}
-	return nil
-}
diff --git a/relay/meta/relay_meta.go b/relay/meta/relay_meta.go
new file mode 100644
index 00000000..bcbe1045
--- /dev/null
+++ b/relay/meta/relay_meta.go
@@ -0,0 +1,61 @@
+package meta
+
+import (
+	"github.com/gin-gonic/gin"
+	"github.com/songquanpeng/one-api/common/ctxkey"
+	"github.com/songquanpeng/one-api/model"
+	"github.com/songquanpeng/one-api/relay/channeltype"
+	"github.com/songquanpeng/one-api/relay/relaymode"
+	"strings"
+)
+
+type Meta struct {
+	Mode         int
+	ChannelType  int
+	ChannelId    int
+	TokenId      int
+	TokenName    string
+	UserId       int
+	Group        string
+	ModelMapping map[string]string
+	// BaseURL is the proxy url set in the channel config
+	BaseURL  string
+	APIKey   string
+	APIType  int
+	Config   model.ChannelConfig
+	IsStream bool
+	// OriginModelName is the model name from the raw user request
+	OriginModelName string
+	// ActualModelName is the model name after mapping
+	ActualModelName string
+	RequestURLPath  string
+	PromptTokens    int // only for DoResponse
+	SystemPrompt    string
+}
+
+func GetByContext(c *gin.Context) *Meta {
+	meta := Meta{
+		Mode:            relaymode.GetByPath(c.Request.URL.Path),
+		ChannelType:     c.GetInt(ctxkey.Channel),
+		ChannelId:       c.GetInt(ctxkey.ChannelId),
+		TokenId:         c.GetInt(ctxkey.TokenId),
+		TokenName:       c.GetString(ctxkey.TokenName),
+		UserId:          c.GetInt(ctxkey.Id),
+		Group:           c.GetString(ctxkey.Group),
+		ModelMapping:    c.GetStringMapString(ctxkey.ModelMapping),
+		OriginModelName: c.GetString(ctxkey.RequestModel),
+		BaseURL:         c.GetString(ctxkey.BaseURL),
+		APIKey:          strings.TrimPrefix(c.Request.Header.Get("Authorization"), "Bearer "),
+		RequestURLPath:  c.Request.URL.String(),
+		SystemPrompt:    c.GetString(ctxkey.SystemPrompt),
+	}
+	cfg, ok := c.Get(ctxkey.Config)
+	if ok {
+		meta.Config = cfg.(model.ChannelConfig)
+	}
+	if meta.BaseURL == "" {
+		meta.BaseURL = channeltype.ChannelBaseURLs[meta.ChannelType]
+	}
+	meta.APIType = channeltype.ToAPIType(meta.ChannelType)
+	return &meta
+}
diff --git a/relay/model/constant.go b/relay/model/constant.go
index f6cf1924..c9d6d645 100644
--- a/relay/model/constant.go
+++ b/relay/model/constant.go
@@ -1,6 +1,7 @@
 package model
 
 const (
-	ContentTypeText     = "text"
-	ContentTypeImageURL = "image_url"
+	ContentTypeText       = "text"
+	ContentTypeImageURL   = "image_url"
+	ContentTypeInputAudio = "input_audio"
 )
diff --git a/relay/model/general.go b/relay/model/general.go
index fbcc04e8..288c07ff 100644
--- a/relay/model/general.go
+++ b/relay/model/general.go
@@ -1,29 +1,70 @@
 package model
 
 type ResponseFormat struct {
-	Type string `json:"type,omitempty"`
+	Type       string      `json:"type,omitempty"`
+	JsonSchema *JSONSchema `json:"json_schema,omitempty"`
+}
+
+type JSONSchema struct {
+	Description string                 `json:"description,omitempty"`
+	Name        string                 `json:"name"`
+	Schema      map[string]interface{} `json:"schema,omitempty"`
+	Strict      *bool                  `json:"strict,omitempty"`
+}
+
+type Audio struct {
+	Voice  string `json:"voice,omitempty"`
+	Format string `json:"format,omitempty"`
+}
+
+type StreamOptions struct {
+	IncludeUsage bool `json:"include_usage,omitempty"`
 }
 
 type GeneralOpenAIRequest struct {
-	Model            string          `json:"model,omitempty"`
-	Messages         []Message       `json:"messages,omitempty"`
-	Prompt           any             `json:"prompt,omitempty"`
-	Stream           bool            `json:"stream,omitempty"`
-	MaxTokens        int             `json:"max_tokens,omitempty"`
-	Temperature      float64         `json:"temperature,omitempty"`
-	TopP             float64         `json:"top_p,omitempty"`
-	N                int             `json:"n,omitempty"`
-	Input            any             `json:"input,omitempty"`
-	Instruction      string          `json:"instruction,omitempty"`
-	Size             string          `json:"size,omitempty"`
-	Functions        any             `json:"functions,omitempty"`
-	FrequencyPenalty float64         `json:"frequency_penalty,omitempty"`
-	PresencePenalty  float64         `json:"presence_penalty,omitempty"`
-	ResponseFormat   *ResponseFormat `json:"response_format,omitempty"`
-	Seed             float64         `json:"seed,omitempty"`
-	Tools            any             `json:"tools,omitempty"`
-	ToolChoice       any             `json:"tool_choice,omitempty"`
-	User             string          `json:"user,omitempty"`
+	// https://platform.openai.com/docs/api-reference/chat/create
+	Messages            []Message       `json:"messages,omitempty"`
+	Model               string          `json:"model,omitempty"`
+	Store               *bool           `json:"store,omitempty"`
+	Metadata            any             `json:"metadata,omitempty"`
+	FrequencyPenalty    *float64        `json:"frequency_penalty,omitempty"`
+	LogitBias           any             `json:"logit_bias,omitempty"`
+	Logprobs            *bool           `json:"logprobs,omitempty"`
+	TopLogprobs         *int            `json:"top_logprobs,omitempty"`
+	MaxTokens           int             `json:"max_tokens,omitempty"`
+	MaxCompletionTokens *int            `json:"max_completion_tokens,omitempty"`
+	N                   int             `json:"n,omitempty"`
+	Modalities          []string        `json:"modalities,omitempty"`
+	Prediction          any             `json:"prediction,omitempty"`
+	Audio               *Audio          `json:"audio,omitempty"`
+	PresencePenalty     *float64        `json:"presence_penalty,omitempty"`
+	ResponseFormat      *ResponseFormat `json:"response_format,omitempty"`
+	Seed                float64         `json:"seed,omitempty"`
+	ServiceTier         *string         `json:"service_tier,omitempty"`
+	Stop                any             `json:"stop,omitempty"`
+	Stream              bool            `json:"stream,omitempty"`
+	StreamOptions       *StreamOptions  `json:"stream_options,omitempty"`
+	Temperature         *float64        `json:"temperature,omitempty"`
+	TopP                *float64        `json:"top_p,omitempty"`
+	TopK                int             `json:"top_k,omitempty"`
+	Tools               []Tool          `json:"tools,omitempty"`
+	ToolChoice          any             `json:"tool_choice,omitempty"`
+	ParallelTooCalls    *bool           `json:"parallel_tool_calls,omitempty"`
+	User                string          `json:"user,omitempty"`
+	FunctionCall        any             `json:"function_call,omitempty"`
+	Functions           any             `json:"functions,omitempty"`
+	// https://platform.openai.com/docs/api-reference/embeddings/create
+	Input          any    `json:"input,omitempty"`
+	EncodingFormat string `json:"encoding_format,omitempty"`
+	Dimensions     int    `json:"dimensions,omitempty"`
+	// https://platform.openai.com/docs/api-reference/images/create
+	Prompt  any     `json:"prompt,omitempty"`
+	Quality *string `json:"quality,omitempty"`
+	Size    string  `json:"size,omitempty"`
+	Style   *string `json:"style,omitempty"`
+	// Others
+	Instruction string `json:"instruction,omitempty"`
+	NumCtx      int    `json:"num_ctx,omitempty"`
 }
 
 func (r GeneralOpenAIRequest) ParseInput() []string {
diff --git a/relay/model/image.go b/relay/model/image.go
new file mode 100644
index 00000000..bab84256
--- /dev/null
+++ b/relay/model/image.go
@@ -0,0 +1,12 @@
+package model
+
+type ImageRequest struct {
+	Model          string `json:"model"`
+	Prompt         string `json:"prompt" binding:"required"`
+	N              int    `json:"n,omitempty"`
+	Size           string `json:"size,omitempty"`
+	Quality        string `json:"quality,omitempty"`
+	ResponseFormat string `json:"response_format,omitempty"`
+	Style          string `json:"style,omitempty"`
+	User           string `json:"user,omitempty"`
+}
diff --git a/relay/model/message.go b/relay/model/message.go
index c6c8a271..b908f989 100644
--- a/relay/model/message.go
+++ b/relay/model/message.go
@@ -1,9 +1,11 @@
 package model
 
 type Message struct {
-	Role    string  `json:"role"`
-	Content any     `json:"content"`
-	Name    *string `json:"name,omitempty"`
+	Role       string  `json:"role,omitempty"`
+	Content    any     `json:"content,omitempty"`
+	Name       *string `json:"name,omitempty"`
+	ToolCalls  []Tool  `json:"tool_calls,omitempty"`
+	ToolCallId string  `json:"tool_call_id,omitempty"`
 }
 
 func (m Message) IsStringContent() bool {
diff --git a/relay/model/tool.go b/relay/model/tool.go
new file mode 100644
index 00000000..75dbb8f7
--- /dev/null
+++ b/relay/model/tool.go
@@ -0,0 +1,14 @@
+package model
+
+type Tool struct {
+	Id       string   `json:"id,omitempty"`
+	Type     string   `json:"type,omitempty"` // when splicing claude tools stream messages, it is empty
+	Function Function `json:"function"`
+}
+
+type Function struct {
+	Description string `json:"description,omitempty"`
+	Name        string `json:"name,omitempty"`       // when splicing claude tools stream messages, it is empty
+	Parameters  any    `json:"parameters,omitempty"` // request
+	Arguments   any    `json:"arguments,omitempty"`  // response
+}
diff --git a/relay/relaymode/define.go b/relay/relaymode/define.go
new file mode 100644
index 00000000..aa771205
--- /dev/null
+++ b/relay/relaymode/define.go
@@ -0,0 +1,16 @@
+package relaymode
+
+const (
+	Unknown = iota
+	ChatCompletions
+	Completions
+	Embeddings
+	Moderations
+	ImagesGenerations
+	Edits
+	AudioSpeech
+	AudioTranscription
+	AudioTranslation
+	// Proxy is a special relay mode for proxying requests to custom upstream
+	Proxy
+)
diff --git a/relay/relaymode/helper.go b/relay/relaymode/helper.go
new file mode 100644
index 00000000..2cde5b85
--- /dev/null
+++ b/relay/relaymode/helper.go
@@ -0,0 +1,31 @@
+package relaymode
+
+import "strings"
+
+func GetByPath(path string) int {
+	relayMode := Unknown
+	if strings.HasPrefix(path, "/v1/chat/completions") {
+		relayMode = ChatCompletions
+	} else if strings.HasPrefix(path, "/v1/completions") {
+		relayMode = Completions
+	} else if strings.HasPrefix(path, "/v1/embeddings") {
+		relayMode = Embeddings
+	} else if strings.HasSuffix(path, "embeddings") {
+		relayMode = Embeddings
+	} else if strings.HasPrefix(path, "/v1/moderations") {
+		relayMode = Moderations
+	} else if strings.HasPrefix(path, "/v1/images/generations") {
+		relayMode = ImagesGenerations
+	} else if strings.HasPrefix(path, "/v1/edits") {
+		relayMode = Edits
+	} else if strings.HasPrefix(path, "/v1/audio/speech") {
+		relayMode = AudioSpeech
+	} else if strings.HasPrefix(path, "/v1/audio/transcriptions") {
+		relayMode = AudioTranscription
+	} else if strings.HasPrefix(path, "/v1/audio/translations") {
+		relayMode = AudioTranslation
+	} else if strings.HasPrefix(path, "/v1/oneapi/proxy") {
+		relayMode = Proxy
+	}
+	return relayMode
+}
diff --git a/relay/util/billing.go b/relay/util/billing.go
deleted file mode 100644
index 495d011e..00000000
--- a/relay/util/billing.go
+++ /dev/null
@@ -1,19 +0,0 @@
-package util
-
-import (
-	"context"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/model"
-)
-
-func ReturnPreConsumedQuota(ctx context.Context, preConsumedQuota int64, tokenId int) {
-	if preConsumedQuota != 0 {
-		go func(ctx context.Context) {
-			// return pre-consumed quota
-			err := model.PostConsumeTokenQuota(tokenId, -preConsumedQuota)
-			if err != nil {
-				logger.Error(ctx, "error return pre-consumed quota: "+err.Error())
-			}
-		}(ctx)
-	}
-}
diff --git a/relay/util/common.go b/relay/util/common.go
deleted file mode 100644
index 535ef680..00000000
--- a/relay/util/common.go
+++ /dev/null
@@ -1,187 +0,0 @@
-package util
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/common/logger"
-	"github.com/songquanpeng/one-api/model"
-	relaymodel "github.com/songquanpeng/one-api/relay/model"
-	"io"
-	"net/http"
-	"strconv"
-	"strings"
-
-	"github.com/gin-gonic/gin"
-)
-
-func ShouldDisableChannel(err *relaymodel.Error, statusCode int) bool {
-	if !config.AutomaticDisableChannelEnabled {
-		return false
-	}
-	if err == nil {
-		return false
-	}
-	if statusCode == http.StatusUnauthorized {
-		return true
-	}
-	switch err.Type {
-	case "insufficient_quota":
-		return true
-	// https://docs.anthropic.com/claude/reference/errors
-	case "authentication_error":
-		return true
-	case "permission_error":
-		return true
-	case "forbidden":
-		return true
-	}
-	if err.Code == "invalid_api_key" || err.Code == "account_deactivated" {
-		return true
-	}
-	if strings.HasPrefix(err.Message, "Your credit balance is too low") { // anthropic
-		return true
-	} else if strings.HasPrefix(err.Message, "This organization has been disabled.") {
-		return true
-	}
-	return false
-}
-
-func ShouldEnableChannel(err error, openAIErr *relaymodel.Error) bool {
-	if !config.AutomaticEnableChannelEnabled {
-		return false
-	}
-	if err != nil {
-		return false
-	}
-	if openAIErr != nil {
-		return false
-	}
-	return true
-}
-
-type GeneralErrorResponse struct {
-	Error    relaymodel.Error `json:"error"`
-	Message  string           `json:"message"`
-	Msg      string           `json:"msg"`
-	Err      string           `json:"err"`
-	ErrorMsg string           `json:"error_msg"`
-	Header   struct {
-		Message string `json:"message"`
-	} `json:"header"`
-	Response struct {
-		Error struct {
-			Message string `json:"message"`
-		} `json:"error"`
-	} `json:"response"`
-}
-
-func (e GeneralErrorResponse) ToMessage() string {
-	if e.Error.Message != "" {
-		return e.Error.Message
-	}
-	if e.Message != "" {
-		return e.Message
-	}
-	if e.Msg != "" {
-		return e.Msg
-	}
-	if e.Err != "" {
-		return e.Err
-	}
-	if e.ErrorMsg != "" {
-		return e.ErrorMsg
-	}
-	if e.Header.Message != "" {
-		return e.Header.Message
-	}
-	if e.Response.Error.Message != "" {
-		return e.Response.Error.Message
-	}
-	return ""
-}
-
-func RelayErrorHandler(resp *http.Response) (ErrorWithStatusCode *relaymodel.ErrorWithStatusCode) {
-	ErrorWithStatusCode = &relaymodel.ErrorWithStatusCode{
-		StatusCode: resp.StatusCode,
-		Error: relaymodel.Error{
-			Message: "",
-			Type:    "upstream_error",
-			Code:    "bad_response_status_code",
-			Param:   strconv.Itoa(resp.StatusCode),
-		},
-	}
-	responseBody, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return
-	}
-	if config.DebugEnabled {
-		logger.SysLog(fmt.Sprintf("error happened, status code: %d, response: \n%s", resp.StatusCode, string(responseBody)))
-	}
-	err = resp.Body.Close()
-	if err != nil {
-		return
-	}
-	var errResponse GeneralErrorResponse
-	err = json.Unmarshal(responseBody, &errResponse)
-	if err != nil {
-		return
-	}
-	if errResponse.Error.Message != "" {
-		// OpenAI format error, so we override the default one
-		ErrorWithStatusCode.Error = errResponse.Error
-	} else {
-		ErrorWithStatusCode.Error.Message = errResponse.ToMessage()
-	}
-	if ErrorWithStatusCode.Error.Message == "" {
-		ErrorWithStatusCode.Error.Message = fmt.Sprintf("bad response status code %d", resp.StatusCode)
-	}
-	return
-}
-
-func GetFullRequestURL(baseURL string, requestURL string, channelType int) string {
-	fullRequestURL := fmt.Sprintf("%s%s", baseURL, requestURL)
-
-	if strings.HasPrefix(baseURL, "https://gateway.ai.cloudflare.com") {
-		switch channelType {
-		case common.ChannelTypeOpenAI:
-			fullRequestURL = fmt.Sprintf("%s%s", baseURL, strings.TrimPrefix(requestURL, "/v1"))
-		case common.ChannelTypeAzure:
-			fullRequestURL = fmt.Sprintf("%s%s", baseURL, strings.TrimPrefix(requestURL, "/openai/deployments"))
-		}
-	}
-	return fullRequestURL
-}
-
-func PostConsumeQuota(ctx context.Context, tokenId int, quotaDelta int64, totalQuota int64, userId int, channelId int, modelRatio float64, groupRatio float64, modelName string, tokenName string) {
-	// quotaDelta is remaining quota to be consumed
-	err := model.PostConsumeTokenQuota(tokenId, quotaDelta)
-	if err != nil {
-		logger.SysError("error consuming token remain quota: " + err.Error())
-	}
-	err = model.CacheUpdateUserQuota(ctx, userId)
-	if err != nil {
-		logger.SysError("error update user quota cache: " + err.Error())
-	}
-	// totalQuota is total quota consumed
-	if totalQuota != 0 {
-		logContent := fmt.Sprintf("模型倍率 %.2f，分组倍率 %.2f", modelRatio, groupRatio)
-		model.RecordConsumeLog(ctx, userId, channelId, int(totalQuota), 0, modelName, tokenName, totalQuota, logContent)
-		model.UpdateUserUsedQuotaAndRequestCount(userId, totalQuota)
-		model.UpdateChannelUsedQuota(channelId, totalQuota)
-	}
-	if totalQuota <= 0 {
-		logger.Error(ctx, fmt.Sprintf("totalQuota consumed is %d, something is wrong", totalQuota))
-	}
-}
-
-func GetAzureAPIVersion(c *gin.Context) string {
-	query := c.Request.URL.Query()
-	apiVersion := query.Get("api-version")
-	if apiVersion == "" {
-		apiVersion = c.GetString(common.ConfigKeyAPIVersion)
-	}
-	return apiVersion
-}
diff --git a/relay/util/init.go b/relay/util/init.go
deleted file mode 100644
index 03dad31b..00000000
--- a/relay/util/init.go
+++ /dev/null
@@ -1,24 +0,0 @@
-package util
-
-import (
-	"github.com/songquanpeng/one-api/common/config"
-	"net/http"
-	"time"
-)
-
-var HTTPClient *http.Client
-var ImpatientHTTPClient *http.Client
-
-func init() {
-	if config.RelayTimeout == 0 {
-		HTTPClient = &http.Client{}
-	} else {
-		HTTPClient = &http.Client{
-			Timeout: time.Duration(config.RelayTimeout) * time.Second,
-		}
-	}
-
-	ImpatientHTTPClient = &http.Client{
-		Timeout: 5 * time.Second,
-	}
-}
diff --git a/relay/util/model_mapping.go b/relay/util/model_mapping.go
deleted file mode 100644
index 39e062a1..00000000
--- a/relay/util/model_mapping.go
+++ /dev/null
@@ -1,12 +0,0 @@
-package util
-
-func GetMappedModelName(modelName string, mapping map[string]string) (string, bool) {
-	if mapping == nil {
-		return modelName, false
-	}
-	mappedModelName := mapping[modelName]
-	if mappedModelName != "" {
-		return mappedModelName, true
-	}
-	return modelName, false
-}
diff --git a/relay/util/relay_meta.go b/relay/util/relay_meta.go
deleted file mode 100644
index 31b9d2b4..00000000
--- a/relay/util/relay_meta.go
+++ /dev/null
@@ -1,55 +0,0 @@
-package util
-
-import (
-	"github.com/gin-gonic/gin"
-	"github.com/songquanpeng/one-api/common"
-	"github.com/songquanpeng/one-api/relay/constant"
-	"strings"
-)
-
-type RelayMeta struct {
-	Mode            int
-	ChannelType     int
-	ChannelId       int
-	TokenId         int
-	TokenName       string
-	UserId          int
-	Group           string
-	ModelMapping    map[string]string
-	BaseURL         string
-	APIVersion      string
-	APIKey          string
-	APIType         int
-	Config          map[string]string
-	IsStream        bool
-	OriginModelName string
-	ActualModelName string
-	RequestURLPath  string
-	PromptTokens    int // only for DoResponse
-}
-
-func GetRelayMeta(c *gin.Context) *RelayMeta {
-	meta := RelayMeta{
-		Mode:           constant.Path2RelayMode(c.Request.URL.Path),
-		ChannelType:    c.GetInt("channel"),
-		ChannelId:      c.GetInt("channel_id"),
-		TokenId:        c.GetInt("token_id"),
-		TokenName:      c.GetString("token_name"),
-		UserId:         c.GetInt("id"),
-		Group:          c.GetString("group"),
-		ModelMapping:   c.GetStringMapString("model_mapping"),
-		BaseURL:        c.GetString("base_url"),
-		APIVersion:     c.GetString(common.ConfigKeyAPIVersion),
-		APIKey:         strings.TrimPrefix(c.Request.Header.Get("Authorization"), "Bearer "),
-		Config:         nil,
-		RequestURLPath: c.Request.URL.String(),
-	}
-	if meta.ChannelType == common.ChannelTypeAzure {
-		meta.APIVersion = GetAzureAPIVersion(c)
-	}
-	if meta.BaseURL == "" {
-		meta.BaseURL = common.ChannelBaseURLs[meta.ChannelType]
-	}
-	meta.APIType = constant.ChannelType2APIType(meta.ChannelType)
-	return &meta
-}
diff --git a/router/api-router.go b/router/api.go
similarity index 90%
rename from router/api-router.go
rename to router/api.go
index 5b755ede..6d00c6ea 100644
--- a/router/api-router.go
+++ b/router/api.go
@@ -2,6 +2,7 @@ package router
 
 import (
 	"github.com/songquanpeng/one-api/controller"
+	"github.com/songquanpeng/one-api/controller/auth"
 	"github.com/songquanpeng/one-api/middleware"
 
 	"github.com/gin-contrib/gzip"
@@ -21,11 +22,14 @@ func SetApiRouter(router *gin.Engine) {
 		apiRouter.GET("/verification", middleware.CriticalRateLimit(), middleware.TurnstileCheck(), controller.SendEmailVerification)
 		apiRouter.GET("/reset_password", middleware.CriticalRateLimit(), middleware.TurnstileCheck(), controller.SendPasswordResetEmail)
 		apiRouter.POST("/user/reset", middleware.CriticalRateLimit(), controller.ResetPassword)
-		apiRouter.GET("/oauth/github", middleware.CriticalRateLimit(), controller.GitHubOAuth)
-		apiRouter.GET("/oauth/state", middleware.CriticalRateLimit(), controller.GenerateOAuthCode)
-		apiRouter.GET("/oauth/wechat", middleware.CriticalRateLimit(), controller.WeChatAuth)
-		apiRouter.GET("/oauth/wechat/bind", middleware.CriticalRateLimit(), middleware.UserAuth(), controller.WeChatBind)
+		apiRouter.GET("/oauth/github", middleware.CriticalRateLimit(), auth.GitHubOAuth)
+		apiRouter.GET("/oauth/oidc", middleware.CriticalRateLimit(), auth.OidcAuth)
+		apiRouter.GET("/oauth/lark", middleware.CriticalRateLimit(), auth.LarkOAuth)
+		apiRouter.GET("/oauth/state", middleware.CriticalRateLimit(), auth.GenerateOAuthCode)
+		apiRouter.GET("/oauth/wechat", middleware.CriticalRateLimit(), auth.WeChatAuth)
+		apiRouter.GET("/oauth/wechat/bind", middleware.CriticalRateLimit(), middleware.UserAuth(), auth.WeChatBind)
 		apiRouter.GET("/oauth/email/bind", middleware.CriticalRateLimit(), middleware.UserAuth(), controller.EmailBind)
+		apiRouter.POST("/topup", middleware.AdminAuth(), controller.AdminTopUp)
 
 		userRoute := apiRouter.Group("/user")
 		{
@@ -43,6 +47,7 @@ func SetApiRouter(router *gin.Engine) {
 				selfRoute.GET("/token", controller.GenerateAccessToken)
 				selfRoute.GET("/aff", controller.GetAffCode)
 				selfRoute.POST("/topup", controller.TopUp)
+				selfRoute.GET("/available_models", controller.GetUserAvailableModels)
 			}
 
 			adminRoute := userRoute.Group("/")
@@ -68,7 +73,7 @@ func SetApiRouter(router *gin.Engine) {
 		{
 			channelRoute.GET("/", controller.GetAllChannels)
 			channelRoute.GET("/search", controller.SearchChannels)
-			channelRoute.GET("/models", controller.ListModels)
+			channelRoute.GET("/models", controller.ListAllModels)
 			channelRoute.GET("/:id", controller.GetChannel)
 			channelRoute.GET("/test", controller.TestChannels)
 			channelRoute.GET("/test/:id", controller.TestChannel)
diff --git a/router/relay-router.go b/router/relay.go
similarity index 97%
rename from router/relay-router.go
rename to router/relay.go
index 65072c86..8f3c7303 100644
--- a/router/relay-router.go
+++ b/router/relay.go
@@ -9,6 +9,7 @@ import (
 
 func SetRelayRouter(router *gin.Engine) {
 	router.Use(middleware.CORS())
+	router.Use(middleware.GzipDecodeMiddleware())
 	// https://platform.openai.com/docs/api-reference/introduction
 	modelsRouter := router.Group("/v1/models")
 	modelsRouter.Use(middleware.TokenAuth())
@@ -19,6 +20,7 @@ func SetRelayRouter(router *gin.Engine) {
 	relayV1Router := router.Group("/v1")
 	relayV1Router.Use(middleware.RelayPanicRecover(), middleware.TokenAuth(), middleware.Distribute())
 	{
+		relayV1Router.Any("/oneapi/proxy/:channelid/*target", controller.Relay)
 		relayV1Router.POST("/completions", controller.Relay)
 		relayV1Router.POST("/chat/completions", controller.Relay)
 		relayV1Router.POST("/edits", controller.Relay)
diff --git a/router/web-router.go b/router/web.go
similarity index 100%
rename from router/web-router.go
rename to router/web.go
diff --git a/web/README.md b/web/README.md
index 29f4713e..829271e2 100644
--- a/web/README.md
+++ b/web/README.md
@@ -2,6 +2,9 @@
 
 > 每个文件夹代表一个主题，欢迎提交你的主题
 
+> [!WARNING]
+> 不是每一个主题都及时同步了所有功能，由于精力有限，优先更新默认主题，其他主题欢迎 & 期待 PR
+
 ## 提交新的主题
 
 > 欢迎在页面底部保留你和 One API 的版权信息以及指向链接
diff --git a/web/air/public/index.html b/web/air/public/index.html
index 36365c7e..e0de002d 100644
--- a/web/air/public/index.html
+++ b/web/air/public/index.html
@@ -1,5 +1,5 @@
 <!DOCTYPE html>
-<html lang="en">
+<html lang="zh-CN">
 <head>
   <meta charset="utf-8" />
   <link rel="icon" href="logo.png" />
diff --git a/web/air/src/components/ChannelsTable.js b/web/air/src/components/ChannelsTable.js
index dee21a01..c384d50c 100644
--- a/web/air/src/components/ChannelsTable.js
+++ b/web/air/src/components/ChannelsTable.js
@@ -437,7 +437,7 @@ const ChannelsTable = () => {
     if (success) {
       record.response_time = time * 1000;
       record.test_time = Date.now() / 1000;
-      showInfo(`通道 ${record.name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
+      showInfo(`渠道 ${record.name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
     } else {
       showError(message);
     }
@@ -447,7 +447,7 @@ const ChannelsTable = () => {
     const res = await API.get(`/api/channel/test?scope=${scope}`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已成功开始测试通道，请刷新页面查看结果。');
+      showInfo('已成功开始测试渠道，请刷新页面查看结果。');
     } else {
       showError(message);
     }
@@ -470,7 +470,7 @@ const ChannelsTable = () => {
     if (success) {
       record.balance = balance;
       record.balance_updated_time = Date.now() / 1000;
-      showInfo(`通道 ${record.name} 余额更新成功！`);
+      showInfo(`渠道 ${record.name} 余额更新成功！`);
     } else {
       showError(message);
     }
@@ -481,7 +481,7 @@ const ChannelsTable = () => {
     const res = await API.get(`/api/channel/update_balance`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已更新完毕所有已启用通道余额！');
+      showInfo('已更新完毕所有已启用渠道余额！');
     } else {
       showError(message);
     }
@@ -490,7 +490,7 @@ const ChannelsTable = () => {
 
   const batchDeleteChannels = async () => {
     if (selectedChannels.length === 0) {
-      showError('请先选择要删除的通道！');
+      showError('请先选择要删除的渠道！');
       return;
     }
     setLoading(true);
@@ -501,7 +501,7 @@ const ChannelsTable = () => {
     const res = await API.post(`/api/channel/batch`, { ids: ids });
     const { success, message, data } = res.data;
     if (success) {
-      showSuccess(`已删除 ${data} 个通道！`);
+      showSuccess(`已删除 ${data} 个渠道！`);
       await refresh();
     } else {
       showError(message);
@@ -513,7 +513,7 @@ const ChannelsTable = () => {
     const res = await API.post(`/api/channel/fix`);
     const { success, message, data } = res.data;
     if (success) {
-      showSuccess(`已修复 ${data} 个通道！`);
+      showSuccess(`已修复 ${data} 个渠道！`);
       await refresh();
     } else {
       showError(message);
@@ -633,7 +633,7 @@ const ChannelsTable = () => {
               onConfirm={() => { testChannels("all") }}
               position={isMobile() ? 'top' : 'left'}
             >
-              <Button theme="light" type="warning" style={{ marginRight: 8 }}>测试所有通道</Button>
+              <Button theme="light" type="warning" style={{ marginRight: 8 }}>测试所有渠道</Button>
             </Popconfirm>
             <Popconfirm
               title="确定？"
@@ -648,16 +648,16 @@ const ChannelsTable = () => {
             okType={'secondary'}
             onConfirm={updateAllChannelsBalance}
           >
-            <Button theme="light" type="secondary" style={{ marginRight: 8 }}>更新所有已启用通道余额</Button>
+            <Button theme="light" type="secondary" style={{ marginRight: 8 }}>更新所有已启用渠道余额</Button>
           </Popconfirm> */}
             <Popconfirm
-              title="确定是否要删除禁用通道？"
+              title="确定是否要删除禁用渠道？"
               content="此修改将不可逆"
               okType={'danger'}
               onConfirm={deleteAllDisabledChannels}
               position={isMobile() ? 'top' : 'left'}
             >
-              <Button theme="light" type="danger" style={{ marginRight: 8 }}>删除禁用通道</Button>
+              <Button theme="light" type="danger" style={{ marginRight: 8 }}>删除禁用渠道</Button>
             </Popconfirm>
 
             <Button theme="light" type="primary" style={{ marginRight: 8 }} onClick={refresh}>刷新</Button>
@@ -673,7 +673,7 @@ const ChannelsTable = () => {
               setEnableBatchDelete(v);
             }}></Switch>
             <Popconfirm
-              title="确定是否要删除所选通道？"
+              title="确定是否要删除所选渠道？"
               content="此修改将不可逆"
               okType={'danger'}
               onConfirm={batchDeleteChannels}
@@ -681,7 +681,7 @@ const ChannelsTable = () => {
               position={'top'}
             >
               <Button disabled={!enableBatchDelete} theme="light" type="danger"
-                style={{ marginRight: 8 }}>删除所选通道</Button>
+                style={{ marginRight: 8 }}>删除所选渠道</Button>
             </Popconfirm>
             <Popconfirm
               title="确定是否要修复数据库一致性？"
diff --git a/web/air/src/components/OperationSetting.js b/web/air/src/components/OperationSetting.js
index b823bb28..6356ac66 100644
--- a/web/air/src/components/OperationSetting.js
+++ b/web/air/src/components/OperationSetting.js
@@ -261,7 +261,7 @@ const OperationSetting = () => {
               value={inputs.ChannelDisableThreshold}
               type='number'
               min='0'
-              placeholder='单位秒，当运行通道全部测试时，超过此时间将自动禁用通道'
+              placeholder='单位秒，当运行渠道全部测试时，超过此时间将自动禁用渠道'
             />
             <Form.Input
               label='额度提醒阈值'
@@ -277,13 +277,13 @@ const OperationSetting = () => {
           <Form.Group inline>
             <Form.Checkbox
               checked={inputs.AutomaticDisableChannelEnabled === 'true'}
-              label='失败时自动禁用通道'
+              label='失败时自动禁用渠道'
               name='AutomaticDisableChannelEnabled'
               onChange={handleInputChange}
             />
             <Form.Checkbox
               checked={inputs.AutomaticEnableChannelEnabled === 'true'}
-              label='成功时自动启用通道'
+              label='成功时自动启用渠道'
               name='AutomaticEnableChannelEnabled'
               onChange={handleInputChange}
             />
diff --git a/web/air/src/components/PersonalSetting.js b/web/air/src/components/PersonalSetting.js
index 45a5b776..ef4acf14 100644
--- a/web/air/src/components/PersonalSetting.js
+++ b/web/air/src/components/PersonalSetting.js
@@ -47,7 +47,7 @@ const PersonalSetting = () => {
   const [countdown, setCountdown] = useState(30);
   const [affLink, setAffLink] = useState('');
   const [systemToken, setSystemToken] = useState('');
-  // const [models, setModels] = useState([]);
+  const [models, setModels] = useState([]);
   const [openTransfer, setOpenTransfer] = useState(false);
   const [transferAmount, setTransferAmount] = useState(0);
 
@@ -72,7 +72,7 @@ const PersonalSetting = () => {
         console.log(userState);
       }
     );
-    // loadModels().then();
+    loadModels().then();
     getAffLink().then();
     setTransferAmount(getQuotaPerUnit());
   }, []);
@@ -127,16 +127,16 @@ const PersonalSetting = () => {
     }
   };
 
-  // const loadModels = async () => {
-  //   let res = await API.get(`/api/user/models`);
-  //   const { success, message, data } = res.data;
-  //   if (success) {
-  //     setModels(data);
-  //     console.log(data);
-  //   } else {
-  //     showError(message);
-  //   }
-  // };
+  const loadModels = async () => {
+    let res = await API.get(`/api/user/available_models`);
+    const { success, message, data } = res.data;
+    if (success) {
+      setModels(data);
+      console.log(data);
+    } else {
+      showError(message);
+    }
+  };
 
   const handleAffLinkClick = async (e) => {
     e.target.select();
@@ -344,7 +344,7 @@ const PersonalSetting = () => {
               }
             >
               <Typography.Title heading={6}>调用信息</Typography.Title>
-              {/* <Typography.Title heading={6}>可用模型</Typography.Title>
+              <p>可用模型（可点击复制）</p>
               <div style={{ marginTop: 10 }}>
                 <Space wrap>
                   {models.map((model) => (
@@ -355,7 +355,7 @@ const PersonalSetting = () => {
                     </Tag>
                   ))}
                 </Space>
-              </div> */}
+              </div>
             </Card>
             {/* <Card
               footer={
diff --git a/web/air/src/components/TokensTable.js b/web/air/src/components/TokensTable.js
index c106b388..48836c85 100644
--- a/web/air/src/components/TokensTable.js
+++ b/web/air/src/components/TokensTable.js
@@ -11,12 +11,14 @@ import EditToken from '../pages/Token/EditToken';
 const COPY_OPTIONS = [
   { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'ChatGPT Web & Midjourney', value: 'ama' },
-  { key: 'opencat', text: 'OpenCat', value: 'opencat' }
+  { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 const OPEN_LINK_OPTIONS = [
   { key: 'ama', text: 'ChatGPT Web & Midjourney', value: 'ama' },
-  { key: 'opencat', text: 'OpenCat', value: 'opencat' }
+  { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' }
 ];
 
 function renderTimestamp(timestamp) {
@@ -60,7 +62,12 @@ const TokensTable = () => {
         onOpenLink('next-mj');
       }
     },
-    { node: 'item', key: 'opencat', name: 'OpenCat', value: 'opencat' }
+    { node: 'item', key: 'opencat', name: 'OpenCat', value: 'opencat' },
+    {
+      node: 'item', key: 'lobechat', name: 'LobeChat', onClick: () => {
+        onOpenLink('lobechat');
+      }
+    }
   ];
 
   const columns = [
@@ -177,6 +184,11 @@ const TokensTable = () => {
                   node: 'item', key: 'opencat', name: 'OpenCat', onClick: () => {
                     onOpenLink('opencat', record.key);
                   }
+                },
+                {
+                  node: 'item', key: 'lobechat', name: 'LobeChat', onClick: () => {
+                    onOpenLink('lobechat');
+                  }
                 }
               ]
             }
@@ -319,7 +331,7 @@ const TokensTable = () => {
     if (nextLink) {
       nextUrl = nextLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     } else {
-      nextUrl = `https://chat.oneapi.pro/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
+      nextUrl = `https://app.nextchat.dev/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     }
 
     let url;
@@ -382,6 +394,9 @@ const TokensTable = () => {
       case 'next-mj':
         url = mjLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
         break;
+      case 'lobechat':
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
+        break;
       default:
         if (!chatLink) {
           showError('管理员未设置聊天链接');
diff --git a/web/air/src/constants/channel.constants.js b/web/air/src/constants/channel.constants.js
index 4bf035f9..e7b25399 100644
--- a/web/air/src/constants/channel.constants.js
+++ b/web/air/src/constants/channel.constants.js
@@ -1,10 +1,13 @@
 export const CHANNEL_OPTIONS = [
   { key: 1, text: 'OpenAI', value: 1, color: 'green' },
   { key: 14, text: 'Anthropic Claude', value: 14, color: 'black' },
+  { key: 33, text: 'AWS', value: 33, color: 'black' },
   { key: 3, text: 'Azure OpenAI', value: 3, color: 'olive' },
   { key: 11, text: 'Google PaLM2', value: 11, color: 'orange' },
   { key: 24, text: 'Google Gemini', value: 24, color: 'orange' },
   { key: 28, text: 'Mistral AI', value: 28, color: 'orange' },
+  { key: 41, text: 'Novita', value: 41, color: 'purple' },
+  { key: 40, text: '字节跳动豆包', value: 40, color: 'blue' },
   { key: 15, text: '百度文心千帆', value: 15, color: 'blue' },
   { key: 17, text: '阿里通义千问', value: 17, color: 'orange' },
   { key: 18, text: '讯飞星火认知', value: 18, color: 'blue' },
@@ -17,6 +20,18 @@ export const CHANNEL_OPTIONS = [
   { key: 29, text: 'Groq', value: 29, color: 'orange' },
   { key: 30, text: 'Ollama', value: 30, color: 'black' },
   { key: 31, text: '零一万物', value: 31, color: 'green' },
+  { key: 32, text: '阶跃星辰', value: 32, color: 'blue' },
+  { key: 34, text: 'Coze', value: 34, color: 'blue' },
+  { key: 35, text: 'Cohere', value: 35, color: 'blue' },
+  { key: 36, text: 'DeepSeek', value: 36, color: 'black' },
+  { key: 37, text: 'Cloudflare', value: 37, color: 'orange' },
+  { key: 38, text: 'DeepL', value: 38, color: 'black' },
+  { key: 39, text: 'together.ai', value: 39, color: 'blue' },
+  { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
+  { key: 43, text: 'Proxy', value: 43, color: 'blue' },
+  { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
+  { key: 45, text: 'xAI', value: 45, color: 'blue' },
+  { key: 46, text: 'Replicate', value: 46, color: 'blue' },
   { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
   { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
   { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },
@@ -34,4 +49,4 @@ export const CHANNEL_OPTIONS = [
 
 for (let i = 0; i < CHANNEL_OPTIONS.length; i++) {
   CHANNEL_OPTIONS[i].label = CHANNEL_OPTIONS[i].text;
-}
\ No newline at end of file
+}
diff --git a/web/air/src/pages/Channel/EditChannel.js b/web/air/src/pages/Channel/EditChannel.js
index efb2cee8..4a810830 100644
--- a/web/air/src/pages/Channel/EditChannel.js
+++ b/web/air/src/pages/Channel/EditChannel.js
@@ -43,6 +43,7 @@ const EditChannel = (props) => {
         base_url: '',
         other: '',
         model_mapping: '',
+        system_prompt: '',
         models: [],
         auto_ban: 1,
         groups: ['default']
@@ -63,7 +64,7 @@ const EditChannel = (props) => {
             let localModels = [];
             switch (value) {
                 case 14:
-                    localModels = ["claude-instant-1.2", "claude-2", "claude-2.0", "claude-2.1", "claude-3-opus-20240229", "claude-3-sonnet-20240229", "claude-3-haiku-20240307"];
+                    localModels = ["claude-instant-1.2", "claude-2", "claude-2.0", "claude-2.1", "claude-3-opus-20240229", "claude-3-sonnet-20240229", "claude-3-haiku-20240307", "claude-3-5-haiku-20241022", "claude-3-5-sonnet-20240620", "claude-3-5-sonnet-20241022"];
                     break;
                 case 11:
                     localModels = ['PaLM-2'];
@@ -78,7 +79,7 @@ const EditChannel = (props) => {
                     localModels = ['chatglm_pro', 'chatglm_std', 'chatglm_lite'];
                     break;
                 case 18:
-                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.5'];
+                    localModels = ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0'];
                     break;
                 case 19:
                     localModels = ['360GPT_S2_V9', 'embedding-bert-512-v1', 'embedding_s1_v1', 'semantic_similarity_s1_v1'];
@@ -304,163 +305,163 @@ const EditChannel = (props) => {
                 width={isMobile() ? '100%' : 600}
             >
                 <Spin spinning={loading}>
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>类型：</Typography.Text>
                     </div>
                     <Select
-                        name='type'
-                        required
-                        optionList={CHANNEL_OPTIONS}
-                        value={inputs.type}
-                        onChange={value => handleInputChange('type', value)}
-                        style={{width: '50%'}}
+                      name='type'
+                      required
+                      optionList={CHANNEL_OPTIONS}
+                      value={inputs.type}
+                      onChange={value => handleInputChange('type', value)}
+                      style={{ width: '50%' }}
                     />
                     {
-                        inputs.type === 3 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Banner type={"warning"} description={
-                                        <>
-                                            注意，<strong>模型部署名称必须和模型名称保持一致</strong>，因为 One API 会把请求体中的
-                                            model
-                                            参数替换为你的部署名称（模型名称中的点会被剔除），<a target='_blank'
-                                                                                              href='https://github.com/songquanpeng/one-api/issues/133?notification_referrer_id=NT_kwDOAmJSYrM2NjIwMzI3NDgyOjM5OTk4MDUw#issuecomment-1571602271'>图片演示</a>。
-                                        </>
-                                    }>
-                                    </Banner>
-                                </div>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>AZURE_OPENAI_ENDPOINT：</Typography.Text>
-                                </div>
-                                <Input
-                                    label='AZURE_OPENAI_ENDPOINT'
-                                    name='azure_base_url'
-                                    placeholder={'请输入 AZURE_OPENAI_ENDPOINT，例如：https://docs-test-001.openai.azure.com'}
-                                    onChange={value => {
-                                        handleInputChange('base_url', value)
-                                    }}
-                                    value={inputs.base_url}
-                                    autoComplete='new-password'
-                                />
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>默认 API 版本：</Typography.Text>
-                                </div>
-                                <Input
-                                    label='默认 API 版本'
-                                    name='azure_other'
-                                    placeholder={'请输入默认 API 版本，例如：2024-03-01-preview，该配置可以被实际的请求查询参数所覆盖'}
-                                    onChange={value => {
-                                        handleInputChange('other', value)
-                                    }}
-                                    value={inputs.other}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
+                      inputs.type === 3 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Banner type={"warning"} description={
+                                    <>
+                                        注意，<strong>模型部署名称必须和模型名称保持一致</strong>，因为 One API 会把请求体中的
+                                        model
+                                        参数替换为你的部署名称（模型名称中的点会被剔除），<a target='_blank'
+                                                                                          href='https://github.com/songquanpeng/one-api/issues/133?notification_referrer_id=NT_kwDOAmJSYrM2NjIwMzI3NDgyOjM5OTk4MDUw#issuecomment-1571602271'>图片演示</a>。
+                                    </>
+                                }>
+                                </Banner>
+                            </div>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>AZURE_OPENAI_ENDPOINT：</Typography.Text>
+                            </div>
+                            <Input
+                              label='AZURE_OPENAI_ENDPOINT'
+                              name='azure_base_url'
+                              placeholder={'请输入 AZURE_OPENAI_ENDPOINT，例如：https://docs-test-001.openai.azure.com'}
+                              onChange={value => {
+                                  handleInputChange('base_url', value)
+                              }}
+                              value={inputs.base_url}
+                              autoComplete='new-password'
+                            />
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>默认 API 版本：</Typography.Text>
+                            </div>
+                            <Input
+                              label='默认 API 版本'
+                              name='azure_other'
+                              placeholder={'请输入默认 API 版本，例如：2024-03-01-preview，该配置可以被实际的请求查询参数所覆盖'}
+                              onChange={value => {
+                                  handleInputChange('other', value)
+                              }}
+                              value={inputs.other}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
                     {
-                        inputs.type === 8 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>Base URL：</Typography.Text>
-                                </div>
-                                <Input
-                                    name='base_url'
-                                    placeholder={'请输入自定义渠道的 Base URL'}
-                                    onChange={value => {
-                                        handleInputChange('base_url', value)
-                                    }}
-                                    value={inputs.base_url}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
+                      inputs.type === 8 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>Base URL：</Typography.Text>
+                            </div>
+                            <Input
+                              name='base_url'
+                              placeholder={'请输入自定义渠道的 Base URL'}
+                              onChange={value => {
+                                  handleInputChange('base_url', value)
+                              }}
+                              value={inputs.base_url}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>名称：</Typography.Text>
                     </div>
                     <Input
-                        required
-                        name='name'
-                        placeholder={'请为渠道命名'}
-                        onChange={value => {
-                            handleInputChange('name', value)
-                        }}
-                        value={inputs.name}
-                        autoComplete='new-password'
+                      required
+                      name='name'
+                      placeholder={'请为渠道命名'}
+                      onChange={value => {
+                          handleInputChange('name', value)
+                      }}
+                      value={inputs.name}
+                      autoComplete='new-password'
                     />
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>分组：</Typography.Text>
                     </div>
                     <Select
-                        placeholder={'请选择可以使用该渠道的分组'}
-                        name='groups'
-                        required
-                        multiple
-                        selection
-                        allowAdditions
-                        additionLabel={'请在系统设置页面编辑分组倍率以添加新的分组：'}
-                        onChange={value => {
-                            handleInputChange('groups', value)
-                        }}
-                        value={inputs.groups}
-                        autoComplete='new-password'
-                        optionList={groupOptions}
+                      placeholder={'请选择可以使用该渠道的分组'}
+                      name='groups'
+                      required
+                      multiple
+                      selection
+                      allowAdditions
+                      additionLabel={'请在系统设置页面编辑分组倍率以添加新的分组：'}
+                      onChange={value => {
+                          handleInputChange('groups', value)
+                      }}
+                      value={inputs.groups}
+                      autoComplete='new-password'
+                      optionList={groupOptions}
                     />
                     {
-                        inputs.type === 18 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>模型版本：</Typography.Text>
-                                </div>
-                                <Input
-                                    name='other'
-                                    placeholder={'请输入星火大模型版本，注意是接口地址中的版本号，例如：v2.1'}
-                                    onChange={value => {
-                                        handleInputChange('other', value)
-                                    }}
-                                    value={inputs.other}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
+                      inputs.type === 18 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>模型版本：</Typography.Text>
+                            </div>
+                            <Input
+                              name='other'
+                              placeholder={'请输入星火大模型版本，注意是接口地址中的版本号，例如：v2.1'}
+                              onChange={value => {
+                                  handleInputChange('other', value)
+                              }}
+                              value={inputs.other}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
                     {
-                        inputs.type === 21 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>知识库 ID：</Typography.Text>
-                                </div>
-                                <Input
-                                    label='知识库 ID'
-                                    name='other'
-                                    placeholder={'请输入知识库 ID，例如：123456'}
-                                    onChange={value => {
-                                        handleInputChange('other', value)
-                                    }}
-                                    value={inputs.other}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
+                      inputs.type === 21 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>知识库 ID：</Typography.Text>
+                            </div>
+                            <Input
+                              label='知识库 ID'
+                              name='other'
+                              placeholder={'请输入知识库 ID，例如：123456'}
+                              onChange={value => {
+                                  handleInputChange('other', value)
+                              }}
+                              value={inputs.other}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>模型：</Typography.Text>
                     </div>
                     <Select
-                        placeholder={'请选择该渠道所支持的模型'}
-                        name='models'
-                        required
-                        multiple
-                        selection
-                        onChange={value => {
-                            handleInputChange('models', value)
-                        }}
-                        value={inputs.models}
-                        autoComplete='new-password'
-                        optionList={modelOptions}
+                      placeholder={'请选择该渠道所支持的模型'}
+                      name='models'
+                      required
+                      multiple
+                      selection
+                      onChange={value => {
+                          handleInputChange('models', value)
+                      }}
+                      value={inputs.models}
+                      autoComplete='new-password'
+                      optionList={modelOptions}
                     />
-                    <div style={{lineHeight: '40px', marginBottom: '12px'}}>
+                    <div style={{ lineHeight: '40px', marginBottom: '12px' }}>
                         <Space>
                             <Button type='primary' onClick={() => {
                                 handleInputChange('models', basicModels);
@@ -473,28 +474,41 @@ const EditChannel = (props) => {
                             }}>清除所有模型</Button>
                         </Space>
                         <Input
-                            addonAfter={
-                                <Button type='primary' onClick={addCustomModel}>填入</Button>
-                            }
-                            placeholder='输入自定义模型名称'
-                            value={customModel}
-                            onChange={(value) => {
-                                setCustomModel(value.trim());
-                            }}
+                          addonAfter={
+                              <Button type='primary' onClick={addCustomModel}>填入</Button>
+                          }
+                          placeholder='输入自定义模型名称'
+                          value={customModel}
+                          onChange={(value) => {
+                              setCustomModel(value.trim());
+                          }}
                         />
                     </div>
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>模型重定向：</Typography.Text>
                     </div>
                     <TextArea
-                        placeholder={`此项可选，用于修改请求体中的模型名称，为一个 JSON 字符串，键为请求中模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
-                        name='model_mapping'
-                        onChange={value => {
-                            handleInputChange('model_mapping', value)
-                        }}
-                        autosize
-                        value={inputs.model_mapping}
-                        autoComplete='new-password'
+                      placeholder={`此项可选，用于修改请求体中的模型名称，为一个 JSON 字符串，键为请求中模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
+                      name='model_mapping'
+                      onChange={value => {
+                          handleInputChange('model_mapping', value)
+                      }}
+                      autosize
+                      value={inputs.model_mapping}
+                      autoComplete='new-password'
+                    />
+                    <div style={{ marginTop: 10 }}>
+                        <Typography.Text strong>系统提示词：</Typography.Text>
+                    </div>
+                    <TextArea
+                      placeholder={`此项可选，用于强制设置给定的系统提示词，请配合自定义模型 & 模型重定向使用，首先创建一个唯一的自定义模型名称并在上面填入，之后将该自定义模型重定向映射到该渠道一个原生支持的模型`}
+                      name='system_prompt'
+                      onChange={value => {
+                          handleInputChange('system_prompt', value)
+                      }}
+                      autosize
+                      value={inputs.system_prompt}
+                      autoComplete='new-password'
                     />
                     <Typography.Text style={{
                         color: 'rgba(var(--semi-blue-5), 1)',
@@ -507,116 +521,116 @@ const EditChannel = (props) => {
                     }>
                         填入模板
                     </Typography.Text>
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>密钥：</Typography.Text>
                     </div>
                     {
                         batch ?
-                            <TextArea
-                                label='密钥'
-                                name='key'
-                                required
-                                placeholder={'请输入密钥，一行一个'}
-                                onChange={value => {
-                                    handleInputChange('key', value)
-                                }}
-                                value={inputs.key}
-                                style={{minHeight: 150, fontFamily: 'JetBrains Mono, Consolas'}}
-                                autoComplete='new-password'
-                            />
-                            :
-                            <Input
-                                label='密钥'
-                                name='key'
-                                required
-                                placeholder={type2secretPrompt(inputs.type)}
-                                onChange={value => {
-                                    handleInputChange('key', value)
-                                }}
-                                value={inputs.key}
-                                autoComplete='new-password'
-                            />
+                          <TextArea
+                            label='密钥'
+                            name='key'
+                            required
+                            placeholder={'请输入密钥，一行一个'}
+                            onChange={value => {
+                                handleInputChange('key', value)
+                            }}
+                            value={inputs.key}
+                            style={{ minHeight: 150, fontFamily: 'JetBrains Mono, Consolas' }}
+                            autoComplete='new-password'
+                          />
+                          :
+                          <Input
+                            label='密钥'
+                            name='key'
+                            required
+                            placeholder={type2secretPrompt(inputs.type)}
+                            onChange={value => {
+                                handleInputChange('key', value)
+                            }}
+                            value={inputs.key}
+                            autoComplete='new-password'
+                          />
                     }
-                    <div style={{marginTop: 10}}>
+                    <div style={{ marginTop: 10 }}>
                         <Typography.Text strong>组织：</Typography.Text>
                     </div>
                     <Input
-                        label='组织，可选，不填则为默认组织'
-                        name='openai_organization'
-                        placeholder='请输入组织org-xxx'
-                        onChange={value => {
-                            handleInputChange('openai_organization', value)
-                        }}
-                        value={inputs.openai_organization}
+                      label='组织，可选，不填则为默认组织'
+                      name='openai_organization'
+                      placeholder='请输入组织org-xxx'
+                      onChange={value => {
+                          handleInputChange('openai_organization', value)
+                      }}
+                      value={inputs.openai_organization}
                     />
-                    <div style={{marginTop: 10, display: 'flex'}}>
+                    <div style={{ marginTop: 10, display: 'flex' }}>
                         <Space>
                             <Checkbox
-                                name='auto_ban'
-                                checked={autoBan}
-                                onChange={
-                                    () => {
-                                        setAutoBan(!autoBan);
-                                    }
-                                }
-                                // onChange={handleInputChange}
+                              name='auto_ban'
+                              checked={autoBan}
+                              onChange={
+                                  () => {
+                                      setAutoBan(!autoBan);
+                                  }
+                              }
+                              // onChange={handleInputChange}
                             />
                             <Typography.Text
-                                strong>是否自动禁用（仅当自动禁用开启时有效），关闭后不会自动禁用该渠道：</Typography.Text>
+                              strong>是否自动禁用（仅当自动禁用开启时有效），关闭后不会自动禁用该渠道：</Typography.Text>
                         </Space>
                     </div>
 
                     {
-                        !isEdit && (
-                            <div style={{marginTop: 10, display: 'flex'}}>
-                                <Space>
-                                    <Checkbox
-                                        checked={batch}
-                                        label='批量创建'
-                                        name='batch'
-                                        onChange={() => setBatch(!batch)}
-                                    />
-                                    <Typography.Text strong>批量创建</Typography.Text>
-                                </Space>
+                      !isEdit && (
+                        <div style={{ marginTop: 10, display: 'flex' }}>
+                            <Space>
+                                <Checkbox
+                                  checked={batch}
+                                  label='批量创建'
+                                  name='batch'
+                                  onChange={() => setBatch(!batch)}
+                                />
+                                <Typography.Text strong>批量创建</Typography.Text>
+                            </Space>
+                        </div>
+                      )
+                    }
+                    {
+                      inputs.type !== 3 && inputs.type !== 8 && inputs.type !== 22 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>代理：</Typography.Text>
                             </div>
-                        )
+                            <Input
+                              label='代理'
+                              name='base_url'
+                              placeholder={'此项可选，用于通过代理站来进行 API 调用'}
+                              onChange={value => {
+                                  handleInputChange('base_url', value)
+                              }}
+                              value={inputs.base_url}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
                     {
-                        inputs.type !== 3 && inputs.type !== 8 && inputs.type !== 22 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>代理：</Typography.Text>
-                                </div>
-                                <Input
-                                    label='代理'
-                                    name='base_url'
-                                    placeholder={'此项可选，用于通过代理站来进行 API 调用'}
-                                    onChange={value => {
-                                        handleInputChange('base_url', value)
-                                    }}
-                                    value={inputs.base_url}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
-                    }
-                    {
-                        inputs.type === 22 && (
-                            <>
-                                <div style={{marginTop: 10}}>
-                                    <Typography.Text strong>私有部署地址：</Typography.Text>
-                                </div>
-                                <Input
-                                    name='base_url'
-                                    placeholder={'请输入私有部署地址，格式为：https://fastgpt.run/api/openapi'}
-                                    onChange={value => {
-                                        handleInputChange('base_url', value)
-                                    }}
-                                    value={inputs.base_url}
-                                    autoComplete='new-password'
-                                />
-                            </>
-                        )
+                      inputs.type === 22 && (
+                        <>
+                            <div style={{ marginTop: 10 }}>
+                                <Typography.Text strong>私有部署地址：</Typography.Text>
+                            </div>
+                            <Input
+                              name='base_url'
+                              placeholder={'请输入私有部署地址，格式为：https://fastgpt.run/api/openapi'}
+                              onChange={value => {
+                                  handleInputChange('base_url', value)
+                              }}
+                              value={inputs.base_url}
+                              autoComplete='new-password'
+                            />
+                        </>
+                      )
                     }
 
                 </Spin>
diff --git a/web/berry/.prettierrc b/web/berry/.prettierrc
new file mode 100644
index 00000000..d5fba07c
--- /dev/null
+++ b/web/berry/.prettierrc
@@ -0,0 +1,8 @@
+{
+  "bracketSpacing": true,
+  "printWidth": 140,
+  "singleQuote": true,
+  "trailingComma": "none",
+  "tabWidth": 2,
+  "useTabs": false
+}
diff --git a/web/berry/package.json b/web/berry/package.json
index f428fd9c..f8265ef7 100644
--- a/web/berry/package.json
+++ b/web/berry/package.json
@@ -15,7 +15,7 @@
     "@mui/utils": "^5.8.6",
     "@mui/x-date-pickers": "^6.18.5",
     "@tabler/icons-react": "^2.44.0",
-    "apexcharts": "^3.35.3",
+    "apexcharts": "3.35.3",
     "axios": "^0.27.2",
     "dayjs": "^1.11.10",
     "formik": "^2.2.9",
@@ -26,7 +26,7 @@
     "notistack": "^3.0.1",
     "prop-types": "^15.8.1",
     "react": "^18.2.0",
-    "react-apexcharts": "^1.4.0",
+    "react-apexcharts": "1.4.0",
     "react-device-detect": "^2.2.2",
     "react-dom": "^18.2.0",
     "react-perfect-scrollbar": "^1.5.8",
diff --git a/web/berry/public/index.html b/web/berry/public/index.html
index 6f232250..9a2457e2 100644
--- a/web/berry/public/index.html
+++ b/web/berry/public/index.html
@@ -1,5 +1,5 @@
 <!DOCTYPE html>
-<html lang="en">
+<html lang="zh-CN">
   <head>
     <title>One API</title>
     <link rel="icon" href="/favicon.ico" />
@@ -11,11 +11,6 @@
       name="description"
       content="OpenAI 接口聚合管理，支持多种渠道包括 Azure，可用于二次分发管理 key，仅单可执行文件，已打包好 Docker 镜像，一键部署，开箱即用"
     />
-    <link rel="preconnect" href="https://fonts.gstatic.com" />
-    <link
-      href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&family=Poppins:wght@400;500;600;700&family=Roboto:wght@400;500;700&display=swap"
-      rel="stylesheet"
-    />
   </head>
 
   <body>
diff --git a/web/berry/src/App.js b/web/berry/src/App.js
index fc54c632..d6422a0f 100644
--- a/web/berry/src/App.js
+++ b/web/berry/src/App.js
@@ -1,8 +1,9 @@
-import { useSelector } from 'react-redux';
+import { useEffect } from 'react';
+import { useSelector, useDispatch } from 'react-redux';
 
 import { ThemeProvider } from '@mui/material/styles';
 import { CssBaseline, StyledEngineProvider } from '@mui/material';
-
+import { SET_THEME } from 'store/actions';
 // routing
 import Routes from 'routes';
 
@@ -20,8 +21,16 @@ import { SnackbarProvider } from 'notistack';
 // ==============================|| APP ||============================== //
 
 const App = () => {
+  const dispatch = useDispatch();
   const customization = useSelector((state) => state.customization);
 
+  useEffect(() => {
+    const storedTheme = localStorage.getItem('theme');
+    if (storedTheme) {
+      dispatch({ type: SET_THEME, theme: storedTheme });
+    }
+  }, [dispatch]);
+
   return (
     <StyledEngineProvider injectFirst>
       <ThemeProvider theme={themes(customization)}>
diff --git a/web/berry/src/assets/fonts/roboto-500.woff2 b/web/berry/src/assets/fonts/roboto-500.woff2
new file mode 100644
index 00000000..2360b721
Binary files /dev/null and b/web/berry/src/assets/fonts/roboto-500.woff2 differ
diff --git a/web/berry/src/assets/fonts/roboto-700.woff2 b/web/berry/src/assets/fonts/roboto-700.woff2
new file mode 100644
index 00000000..4aeda71b
Binary files /dev/null and b/web/berry/src/assets/fonts/roboto-700.woff2 differ
diff --git a/web/berry/src/assets/fonts/roboto-regular.woff2 b/web/berry/src/assets/fonts/roboto-regular.woff2
new file mode 100644
index 00000000..b65a361a
Binary files /dev/null and b/web/berry/src/assets/fonts/roboto-regular.woff2 differ
diff --git a/web/berry/src/assets/images/icons/lark.svg b/web/berry/src/assets/images/icons/lark.svg
new file mode 100644
index 00000000..79688e2a
--- /dev/null
+++ b/web/berry/src/assets/images/icons/lark.svg
@@ -0,0 +1,5 @@
+<svg t="1723134993089" class="icon" viewBox="0 0 1024 1024" version="1.1" xmlns="http://www.w3.org/2000/svg" p-id="7682"
+     width="200" height="200">
+    <path d="M138.67 472.593v267.659l1.085 0.825c30.488 23.11 68.369 45.41 109.072 62.904l1.473 0.63c57.452 24.487 117.23 38.204 176.468 38.39 49.026 0.15 94.773-6.199 138.432-20.266 23.88-7.694 47.048-17.68 69.546-30.137-16.42-1.696-32.855-4.635-49.166-8.6-68.949-16.77-138.975-52.578-203.999-96.138-92.077-61.684-178.453-141.863-242.91-215.267zM597.31 182H350.745l25.776 22.626 6.472 5.7 7.9 6.985 6.658 5.915 4.791 4.275 6.144 5.509 5.17 4.665 4.297 3.899 4.876 4.451 4.077 3.748 3.337 3.085 4.594 4.276 3.238 3.034 3.21 3.026 1.917 1.814 1.633 1.553a1573.085 1573.085 0 0 1 16.368 15.806c61.384 60.116 109.769 118.745 146.283 181.615 34.096-38.136 67.506-67.95 100.368-90.075-0.08-0.224-0.158-0.452-0.233-0.681-9.674-29.434-24.133-63.778-41.697-96.235-20.272-37.46-43.21-69.797-68.284-94.664l-0.331-0.327z m149.927 506.476a42.573 42.573 0 0 1 13.201-4.727c12.04-16.686 22.616-34.537 32.12-53.8 5.835-11.824 11.243-24.101 16.582-37.512l1.185-2.999c0.593-1.509 1.184-3.032 1.776-4.572l1.184-3.099a817.09 817.09 0 0 0 2.963-7.94l0.902-2.466 1.54-4.256 1.24-3.478 0.944-2.668 1.282-3.658 1.32-3.797 1.364-3.963 1.784-5.225 10.079-29.815 1.237-3.615 1.619-4.688 1.193-3.419 1.176-3.34 1.16-3.261 0.766-2.133 1.137-3.139 0.75-2.052 0.747-2.022 1.11-2.977 0.735-1.948 0.731-1.92 0.728-1.893 1.086-2.789 1.08-2.731 0.717-1.79 0.716-1.766 0.715-1.743 0.714-1.721 1.07-2.54 0.713-1.668 0.714-1.649 0.357-0.816 0.715-1.62 0.358-0.803 0.717-1.593 0.72-1.576 0.72-1.56 0.725-1.545 0.727-1.53 0.73-1.515 0.734-1.503 0.739-1.49 0.743-1.478 0.373-0.735 0.75-1.461 0.756-1.452 0.761-1.443 0.768-1.434 0.774-1.426 0.78-1.42 0.789-1.413 0.796-1.407 0.803-1.402 0.813-1.399c0.272-0.465 0.545-0.93 0.82-1.395 6.898-11.644 14.45-22.205 22.804-31.716l0.381-0.432-0.363-0.1a167.736 167.736 0 0 0-17.333-3.808l-1.17-0.187c-19.22-3.015-38.771-2.474-59.435 2.33C766 404.728 707.642 444.655 643.04 520.58c-50.09 58.868-112.372 100.194-178.36 119.035 48.387 29.017 96.32 50.621 141.075 61.507 55.428 13.48 102.617 9.838 140.868-12.298z m40.248 75.25a229.325 229.325 0 0 1-21.023 10.522 420.333 420.333 0 0 1-20.405 18.512c-50.095 42.57-103.865 72.408-161.229 90.891-50.432 16.25-102.815 23.52-158.258 23.355-68.336-0.22-136.434-15.846-201.363-43.52-52.844-22.523-101.532-52.32-138.874-83.06a32.016 32.016 0 0 1-11.662-24.212l-0.003-364.272a663.709 663.709 0 0 1-3.155-4.546l-0.395-0.584c-12.855-19.356-7.842-45.51 11.382-58.72 19.417-13.345 45.98-8.428 59.33 10.98 49.099 71.385 141.145 170.175 245.108 248.023 60.719-0.825 122.443-27.243 174.57-74.197-34.118-63.594-81.951-122.977-145.085-184.807a1497.032 1497.032 0 0 0-16.553-15.971l-3.41-3.23-3.07-2.888-3.727-3.484-3.783-3.514-3.854-3.556-3.938-3.61-5.403-4.921-4.184-3.786-5.043-4.54-4.477-4.008-6.214-5.538-7.36-6.527-9.592-8.463-8.388-7.371-15.995-14.032-10.596-9.324-7.673-6.781-6.151-5.465-4.92-4.397-4.536-4.081-3.615-3.28-3.395-3.104-2.68-2.472-2.058-1.913-2.48-2.324-1.919-1.814-1.871-1.783-2.735-2.632-2.231-2.172-1.92-1.888a702.481 702.481 0 0 1-6.747-6.754c-19.741-20.021-5.898-53.785 21.938-54.458l0.848-0.01h341.332a32 32 0 0 1 21.426 8.232c34.675 31.26 64.87 72.453 90.762 120.299 17.457 32.258 32.066 66.007 42.695 96.357 13.332-5.523 26.586-9.793 39.771-12.858 28.861-6.71 56.686-7.48 83.844-3.22 21.406 3.358 40.992 9.48 62.014 18.392l1.835 0.783 1.842 0.796 1.861 0.816 1.894 0.839 1.94 0.868 1.996 0.903 11.45 5.265c23.934 10.99 25.088 44.57 1.966 57.177-18.753 10.224-33.09 24.764-45.475 45.42l-0.82 1.383-0.65 1.111-0.643 1.119-0.64 1.126-0.634 1.136-0.631 1.144-0.627 1.155-0.623 1.165-0.62 1.176-0.618 1.188-0.615 1.2-0.612 1.214-0.61 1.227-0.609 1.24-0.607 1.257-0.606 1.272-0.605 1.287-0.907 1.963-0.604 1.33-0.605 1.35-0.605 1.368-0.606 1.387-0.606 1.407-0.608 1.428-0.61 1.45-0.919 2.216-0.923 2.267-0.62 1.542-0.622 1.565-0.626 1.591-0.63 1.617-0.633 1.643-0.638 1.67-1.29 3.423-0.981 2.642-0.994 2.71-0.67 1.845-1.017 2.826-1.38 3.88-1.054 2.999-1.072 3.076-1.09 3.155-1.11 3.237-10.062 29.76-1.935 5.673-1.84 5.35-1.42 4.084-1.386 3.945-1.022 2.885-1.35 3.772-0.672 1.862-1.01 2.775c-8.668 23.731-17.234 44.304-27.005 64.106-11.498 23.305-24.457 44.997-39.33 65.303 0.773 15.454-6.924 30.784-21.242 39.07z"
+          p-id="7683" fill="#2c2c2c"></path>
+</svg>
diff --git a/web/berry/src/assets/images/icons/oidc.svg b/web/berry/src/assets/images/icons/oidc.svg
new file mode 100644
index 00000000..96e01f81
--- /dev/null
+++ b/web/berry/src/assets/images/icons/oidc.svg
@@ -0,0 +1,7 @@
+<svg t="1723135116886" class="icon" viewBox="0 0 1024 1024" version="1.1" xmlns="http://www.w3.org/2000/svg"
+     p-id="10969" width="200" height="200">
+    <path d="M512 960C265 960 64 759 64 512S265 64 512 64s448 201 448 448-201 448-448 448z m0-882.6c-239.7 0-434.6 195-434.6 434.6s195 434.6 434.6 434.6 434.6-195 434.6-434.6S751.7 77.4 512 77.4z"
+          p-id="10970" fill="#2c2c2c" stroke="#2c2c2c" stroke-width="60"></path>
+    <path d="M197.7 512c0-78.3 31.6-98.8 87.2-98.8 56.2 0 87.2 20.5 87.2 98.8s-31 98.8-87.2 98.8c-55.7 0-87.2-20.5-87.2-98.8z m130.4 0c0-46.8-7.8-64.5-43.2-64.5-35.2 0-42.9 17.7-42.9 64.5 0 47.1 7.8 63.7 42.9 63.7 35.4 0 43.2-16.6 43.2-63.7zM409.7 415.9h42.1V608h-42.1V415.9zM653.9 512c0 74.2-37.1 96.1-93.6 96.1h-65.9V415.9h65.9c56.5 0 93.6 16.1 93.6 96.1z m-43.5 0c0-49.3-17.7-60.6-52.3-60.6h-21.6v120.7h21.6c35.4 0 52.3-13.3 52.3-60.1zM686.5 512c0-74.2 36.3-98.8 92.7-98.8 18.3 0 33.2 2.2 44.8 6.4v36.3c-11.9-4.2-26-6.6-42.1-6.6-34.6 0-49.8 15.5-49.8 62.6 0 50.1 15.2 62.6 49.3 62.6 15.8 0 30.2-2.2 44.8-7.5v36c-11.3 4.7-28.5 8-46.8 8-56.1-0.2-92.9-18.7-92.9-99z"
+          p-id="10971" fill="#2c2c2c" stroke="#2c2c2c" stroke-width="20"></path>
+</svg>
diff --git a/web/berry/src/assets/images/logo-white.svg b/web/berry/src/assets/images/logo-white.svg
new file mode 100644
index 00000000..d6289b9a
--- /dev/null
+++ b/web/berry/src/assets/images/logo-white.svg
@@ -0,0 +1,13 @@
+<svg viewBox="0 0 590 360" xmlns="http://www.w3.org/2000/svg" version="1.1">
+ <g>
+  <ellipse transform="rotate(103.6 287.6 32.87)" id="svg_1" ry="28.14" rx="28.47" cy="32.87" cx="287.60001" fill="#fff"/>
+  <path id="svg_2" d="m232.92,128.89c3.78,27.29 -1.81,55.44 -17.71,78.09a2.62,2.62 0 0 0 -0.06,2.92c1.24,1.92 2.96,5.05 5.56,4.94q5.25,-0.22 10.79,0.11a1.26,1.26 0 0 1 1.19,1.27l-0.4,42.53a1.31,1.31 0 0 1 -1.31,1.3q-16.77,-0.09 -36.53,0.01q-2.25,0.02 -3.71,-1.56q-16.02,-17.28 -31.98,-35.32c-5.13,-5.8 -10.18,-11.16 -14.86,-17.59a1.35,1.34 -31.1 0 1 0.5,-2q12.88,-6.32 22.13,-17.12q18.18,-21.23 15.08,-48.84q-2.66,-23.7 -22.4,-40.46q-23.43,-19.9 -54.88,-13.86c-4.1,0.79 -7.83,2.5 -11.72,4.12q-11.86,4.94 -20.59,14.64c-14.25,15.81 -20.07,36.4 -15.05,57.16q4.99,20.63 22.86,35.71c10.45,8.81 23.7,13.12 37.26,14.18q1.47,0.11 3.6,2.65c11.68,13.89 24.48,27.72 35.94,41.96a0.43,0.43 0 0 1 -0.21,0.68q-22.51,7.27 -47.37,5.37q-19.4,-1.47 -39.74,-11.22q-18.27,-8.75 -30.59,-21.28q-18.66,-18.98 -28.02,-43.57q-10.8,-28.4 -4.93,-58.67c1.59,-8.17 4.03,-17 7.42,-24.61q5.08,-11.38 11.61,-20.64q25.41,-36.03 68.45,-46.13q32.42,-7.61 64.23,3.92q25.31,9.17 43.2,27.31c16.85,17.09 28.91,40.01 32.24,64z" fill="#fff"/>
+  <path id="svg_3" d="m499.47,180.61c6.45,13.53 16.44,21.75 31.96,22q11.94,0.19 22.17,-5.36q2.21,-1.2 3.93,0.69q12.56,13.78 24.89,28.47q1.21,1.44 1.44,3.13a0.95,0.95 0 0 1 -0.36,0.89c-1.62,1.23 -3.33,2.71 -5.03,3.69q-29.37,17.01 -62.47,11.31c-20.61,-3.55 -39.05,-15.24 -51.47,-32.51q-6.4,-8.89 -9.91,-17.08c-2.62,-6.12 -4.73,-13.3 -5.41,-20.08q-3.96,-39.88 22.94,-67.74c9.48,-9.81 21.15,-16.67 34.39,-19.49c16.54,-3.53 34.64,-1.83 48.77,7.1q13.92,8.79 21.13,20.4q11.07,17.84 10.48,38.92c-0.02,0.94 -0.21,1.81 -0.85,2.54q-7.73,8.77 -18.71,20.16c-1.28,1.32 -2.61,2.26 -4.51,2.23q-24.45,-0.37 -51.64,-0.41q-5.03,0 -10.84,-0.22a0.96,0.95 -11.7 0 0 -0.9,1.36zm1.12,-37.17q-0.55,1.19 -0.63,2.34q-0.08,1.01 0.94,1.03q19.01,0.25 36.98,0.01q0.5,0 0.94,-0.22q0.57,-0.28 0.44,-0.9q-2.34,-11.6 -14.11,-15.25q-3.59,-1.11 -6.44,-0.57q-13.07,2.5 -18.12,13.56z" fill="#fff"/>
+  <path id="svg_4" d="m312.3,100.22a0.5,0.49 -22.1 0 0 0.84,0.35q2.76,-2.64 5.82,-4.31q8.45,-4.62 16.71,-6.57c15.81,-3.72 33.58,-3.2 48.2,3.95q24.49,11.98 35.05,35.76c4.66,10.5 5.44,22.96 5.5,35.35q0.21,49.99 -0.12,88q-0.03,3.06 -0.08,6.16a1.32,1.32 0 0 1 -1.33,1.3q-20.22,-0.18 -40.18,-0.23q-3.64,-0.01 -8.13,-0.44a1.06,1.05 -87.3 0 1 -0.95,-1.05q0.02,-45.49 -0.22,-92.99c-0.03,-6.25 -1.21,-13.88 -5.05,-18.95q-5.33,-7.03 -12.32,-10.18c-10.99,-4.93 -24.52,-1.84 -33.13,6.37q-10.01,9.53 -10.07,23.76q-0.11,25.46 -0.1,48.98c0,3.52 -0.06,8.31 -1.1,11.68c-4.37,14.04 -17.31,19.5 -31.04,16.77c-8.22,-1.64 -15.07,-7.75 -17.62,-15.62q-1.45,-4.49 -1.42,-10.2q0.3,-64.69 0.1,-129.86a0.47,0.47 0 0 1 0.47,-0.47l48.46,-0.35a1.56,1.55 89.4 0 1 1.56,1.54l0.15,11.25z" fill="#fff"/>
+  <path id="svg_5" d="m265.63,344.43a2.02,2.01 76.7 0 0 -1.85,-1.15l-17.03,0.24a2.25,2.22 9.3 0 0 -2.06,1.46l-2.86,7.84a2.47,2.46 -79.1 0 1 -2.38,1.62l-6.23,-0.19q-1.19,-0.04 -0.88,-1.19q1.38,-5.23 2.81,-8.7c3.41,-8.3 6.48,-16.83 10.12,-25.35q2.96,-6.93 5.21,-14.24c0.46,-1.52 1.69,-2.64 3.37,-2.63c2.02,0 4.68,-0.78 5.7,1.58q7.68,17.74 18.16,44.75q0.96,2.46 1.48,5a0.67,0.66 84.3 0 1 -0.65,0.8l-6.05,-0.02q-2.16,-0.01 -3.1,-1.96l-3.76,-7.86zm-16.73,-10.31a0.34,0.34 0 0 0 0.32,0.47l12.85,-0.36a0.34,0.34 0 0 0 0.3,-0.48l-6.84,-14.7a0.34,0.34 0 0 0 -0.62,0.02l-6.01,15.05z" fill="#fff"/>
+  <rect id="svg_6" rx="2.17" height="52.28" width="9.84" y="302.19" x="345.67" fill="#fff"/>
+  <path id="svg_7" d="m303.07,338.46l-0.15,14.42q-0.01,1.55 -1.56,1.52l-5.84,-0.12q-1.79,-0.04 -1.81,-1.83c-0.24,-15.33 -0.25,-30.89 -0.27,-47.22q-0.01,-2.99 2.55,-3.06q12.47,-0.33 20.15,0.8q8.61,1.25 12.86,9.17c2.95,5.49 2.53,13.5 -1.5,18.65c-5.57,7.14 -14.88,6.62 -23.24,6.51a1.17,1.17 0 0 0 -1.19,1.16zm-0.15,-24.81l0.16,12.72a1.72,1.72 0 0 0 1.74,1.7l6.07,-0.08a10.01,7.98 -0.7 0 0 9.91,-8.1l0,-0.2a10.01,7.98 -0.7 0 0 -10.11,-7.86l-6.07,0.08a1.72,1.72 0 0 0 -1.7,1.74z" fill="#fff"/>
+  <rect id="svg_8" rx="3.58" height="7.26" width="79.2" y="322.99" x="107" fill="#fff"/>
+  <rect id="svg_9" rx="3.81" height="7.72" width="79.1" y="322.78" x="417.27" fill="#fff"/>
+ </g>
+</svg>
\ No newline at end of file
diff --git a/web/berry/src/assets/scss/_themes-vars.module.scss b/web/berry/src/assets/scss/_themes-vars.module.scss
index a470b033..661bb6c6 100644
--- a/web/berry/src/assets/scss/_themes-vars.module.scss
+++ b/web/berry/src/assets/scss/_themes-vars.module.scss
@@ -46,11 +46,16 @@ $grey600: #4b5565;
 $grey700: #364152;
 $grey900: #121926;
 
+$tableBackground: #f4f6f8;
+$tableBorderBottom: #f1f3f4;
+
 // ==============================|| DARK THEME VARIANTS ||============================== //
 
 // paper & background
 $darkBackground: #1a223f; // level 3
 $darkPaper: #111936; // level 4
+$darkDivider: rgba(227, 232, 239, 0.2);
+$darkSelectedBack : rgba(124, 77, 255, 0.15);
 
 // dark 800 & 900
 $darkLevel1: #29314f; // level 1
@@ -154,4 +159,9 @@ $darkTextSecondary: #8492c4;
   darkSecondaryDark: $darkSecondaryDark;
   darkSecondary200: $darkSecondary200;
   darkSecondary800: $darkSecondary800;
+
+  darkDivider: $darkDivider;
+  darkSelectedBack: $darkSelectedBack;
+  tableBackground: $tableBackground;
+  tableBorderBottom: $tableBorderBottom;
 }
diff --git a/web/berry/src/assets/scss/fonts.scss b/web/berry/src/assets/scss/fonts.scss
new file mode 100644
index 00000000..c792aab2
--- /dev/null
+++ b/web/berry/src/assets/scss/fonts.scss
@@ -0,0 +1,32 @@
+
+/* roboto-regular */
+@font-face {
+    font-family: 'Roboto';
+    font-style: normal;
+    font-weight: 400;
+    font-display: swap;
+    src: local('Roboto'), url('../fonts/roboto-regular.woff2') format('woff2');
+    unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+2074, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+  }
+
+  /* roboto-500 */
+@font-face {
+    font-family: 'Roboto';
+    font-style: normal;
+    font-weight: 500;
+    font-display: swap;
+    src: local('Roboto'), url('../fonts/roboto-500.woff2') format('woff2');
+    unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+2074, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+  
+    
+/* roboto-700 */
+@font-face {
+    font-family: 'Roboto';
+    font-style: normal;
+    font-weight: 700;
+    font-display: swap;
+    src: local('Roboto'), url('../fonts/roboto-700.woff2') format('woff2');
+    unicode-range: U+0000-00FF, U+0131, U+0152-0153, U+02BB-02BC, U+02C6, U+02DA, U+02DC, U+0304, U+0308, U+0329, U+2000-206F, U+2074, U+20AC, U+2122, U+2191, U+2193, U+2212, U+2215, U+FEFF, U+FFFD;
+}
+    
\ No newline at end of file
diff --git a/web/berry/src/assets/scss/style.scss b/web/berry/src/assets/scss/style.scss
index 17d566e6..5d2d8975 100644
--- a/web/berry/src/assets/scss/style.scss
+++ b/web/berry/src/assets/scss/style.scss
@@ -1,3 +1,4 @@
+@import 'fonts.scss';
 // color variants
 @import 'themes-vars.module.scss';
 
diff --git a/web/berry/src/config.js b/web/berry/src/config.js
index eeeda99a..8c1faf9b 100644
--- a/web/berry/src/config.js
+++ b/web/berry/src/config.js
@@ -22,7 +22,12 @@ const config = {
     turnstile_site_key: '',
     version: '',
     wechat_login: false,
-    wechat_qrcode: ''
+    wechat_qrcode: '',
+    oidc: false,
+    oidc_client_id: '',
+    oidc_authorization_endpoint: '',
+    oidc_token_endpoint: '',
+    oidc_userinfo_endpoint: '',
   }
 };
 
diff --git a/web/berry/src/constants/ChannelConstants.js b/web/berry/src/constants/ChannelConstants.js
index ec049f7d..375adcd9 100644
--- a/web/berry/src/constants/ChannelConstants.js
+++ b/web/berry/src/constants/ChannelConstants.js
@@ -11,6 +11,18 @@ export const CHANNEL_OPTIONS = {
     value: 14,
     color: 'primary'
   },
+  33: {
+    key: 33,
+    text: 'AWS',
+    value: 33,
+    color: 'primary'
+  },
+  37: {
+    key: 37,
+    text: 'Cloudflare',
+    value: 37,
+    color: 'success'
+  },
   3: {
     key: 3,
     text: 'Azure OpenAI',
@@ -35,6 +47,12 @@ export const CHANNEL_OPTIONS = {
     value: 28,
     color: 'warning'
   },
+  40: {
+    key: 40,
+    text: '字节跳动豆包',
+    value: 40,
+    color: 'primary'
+  },
   15: {
     key: 15,
     text: '百度文心千帆',
@@ -107,6 +125,78 @@ export const CHANNEL_OPTIONS = {
     value: 31,
     color: 'primary'
   },
+  32: {
+    key: 32,
+    text: '阶跃星辰',
+    value: 32,
+    color: 'primary'
+  },
+  34: {
+    key: 34,
+    text: 'Coze',
+    value: 34,
+    color: 'primary'
+  },
+  35: {
+    key: 35,
+    text: 'Cohere',
+    value: 35,
+    color: 'primary'
+  },
+  36: {
+    key: 36,
+    text: 'DeepSeek',
+    value: 36,
+    color: 'primary'
+  },
+  38: {
+    key: 38,
+    text: 'DeepL',
+    value: 38,
+    color: 'primary'
+  },
+  39: {
+    key: 39,
+    text: 'together.ai',
+    value: 39,
+    color: 'primary'
+  },
+  42: {
+    key: 42,
+    text: 'VertexAI',
+    value: 42,
+    color: 'primary'
+  },
+  43: {
+    key: 43,
+    text: 'Proxy',
+    value: 43,
+    color: 'primary'
+  },
+  44: {
+    key: 44,
+    text: 'SiliconFlow',
+    value: 44,
+    color: 'primary'
+  },
+  45: {
+    key: 45,
+    text: 'xAI',
+    value: 45,
+    color: 'primary'
+  },
+  45: {
+    key: 46,
+    text: 'Replicate',
+    value: 46,
+    color: 'primary'
+  },
+  41: {
+    key: 41,
+    text: 'Novita',
+    value: 41,
+    color: 'purple'
+  },
   8: {
     key: 8,
     text: '自定义渠道',
diff --git a/web/berry/src/constants/SnackbarConstants.js b/web/berry/src/constants/SnackbarConstants.js
index a05c6652..05f79231 100644
--- a/web/berry/src/constants/SnackbarConstants.js
+++ b/web/berry/src/constants/SnackbarConstants.js
@@ -1,24 +1,56 @@
+import { closeSnackbar } from 'notistack';
+import { IconX } from '@tabler/icons-react';
+import { IconButton } from '@mui/material';
+const action = (snackbarId) => (
+  <>
+    <IconButton
+      onClick={() => {
+        closeSnackbar(snackbarId);
+      }}
+    >
+      <IconX stroke={1.5} size="1.25rem" />
+    </IconButton>
+  </>
+);
+
 export const snackbarConstants = {
   Common: {
     ERROR: {
       variant: 'error',
-      autoHideDuration: 5000
+      autoHideDuration: 5000,
+      preventDuplicate: true,
+      action
     },
     WARNING: {
       variant: 'warning',
-      autoHideDuration: 10000
+      autoHideDuration: 10000,
+      preventDuplicate: true,
+      action
     },
     SUCCESS: {
       variant: 'success',
-      autoHideDuration: 1500
+      autoHideDuration: 1500,
+      preventDuplicate: true,
+      action
     },
     INFO: {
       variant: 'info',
-      autoHideDuration: 3000
+      autoHideDuration: 3000,
+      preventDuplicate: true,
+      action
     },
     NOTICE: {
       variant: 'info',
-      autoHideDuration: 20000
+      autoHideDuration: 20000,
+      preventDuplicate: true,
+      action
+    },
+    COPY: {
+      variant: 'copy',
+      persist: true,
+      preventDuplicate: true,
+      allowDownload: true,
+      action
     }
   },
   Mobile: {
diff --git a/web/berry/src/hooks/useLogin.js b/web/berry/src/hooks/useLogin.js
index 53626577..6d89727d 100644
--- a/web/berry/src/hooks/useLogin.js
+++ b/web/berry/src/hooks/useLogin.js
@@ -48,6 +48,50 @@ const useLogin = () => {
     }
   };
 
+  const larkLogin = async (code, state) => {
+    try {
+      const res = await API.get(`/api/oauth/lark?code=${code}&state=${state}`);
+      const { success, message, data } = res.data;
+      if (success) {
+        if (message === 'bind') {
+          showSuccess('绑定成功！');
+          navigate('/panel');
+        } else {
+          dispatch({ type: LOGIN, payload: data });
+          localStorage.setItem('user', JSON.stringify(data));
+          showSuccess('登录成功！');
+          navigate('/panel');
+        }
+      }
+      return { success, message };
+    } catch (err) {
+      // 请求失败，设置错误信息
+      return { success: false, message: '' };
+    }
+  };
+
+  const oidcLogin = async (code, state) => {
+    try {
+      const res = await API.get(`/api/oauth/oidc?code=${code}&state=${state}`);
+      const { success, message, data } = res.data;
+      if (success) {
+        if (message === 'bind') {
+          showSuccess('绑定成功！');
+          navigate('/panel');
+        } else {
+          dispatch({ type: LOGIN, payload: data });
+          localStorage.setItem('user', JSON.stringify(data));
+          showSuccess('登录成功！');
+          navigate('/panel');
+        }
+      }
+      return { success, message };
+    } catch (err) {
+      // 请求失败，设置错误信息
+      return { success: false, message: '' };
+    }
+  }
+
   const wechatLogin = async (code) => {
     try {
       const res = await API.get(`/api/oauth/wechat?code=${code}`);
@@ -72,7 +116,7 @@ const useLogin = () => {
     navigate('/');
   };
 
-  return { login, logout, githubLogin, wechatLogin };
+  return { login, logout, githubLogin, wechatLogin, larkLogin,oidcLogin };
 };
 
 export default useLogin;
diff --git a/web/berry/src/layout/MainLayout/Header/ProfileSection/index.js b/web/berry/src/layout/MainLayout/Header/ProfileSection/index.js
index 3e351254..e1392dc0 100644
--- a/web/berry/src/layout/MainLayout/Header/ProfileSection/index.js
+++ b/web/berry/src/layout/MainLayout/Header/ProfileSection/index.js
@@ -71,8 +71,8 @@ const ProfileSection = () => {
           alignItems: 'center',
           borderRadius: '27px',
           transition: 'all .2s ease-in-out',
-          borderColor: theme.palette.primary.light,
-          backgroundColor: theme.palette.primary.light,
+          borderColor: theme.typography.menuChip.background,
+          backgroundColor: theme.typography.menuChip.background,
           '&[aria-controls="menu-list-grow"], &:hover': {
             borderColor: theme.palette.primary.main,
             background: `${theme.palette.primary.main}!important`,
diff --git a/web/berry/src/layout/MainLayout/Header/index.js b/web/berry/src/layout/MainLayout/Header/index.js
index 51d40c75..8fd9c950 100644
--- a/web/berry/src/layout/MainLayout/Header/index.js
+++ b/web/berry/src/layout/MainLayout/Header/index.js
@@ -7,6 +7,7 @@ import { Avatar, Box, ButtonBase } from '@mui/material';
 // project imports
 import LogoSection from '../LogoSection';
 import ProfileSection from './ProfileSection';
+import ThemeButton from 'ui-component/ThemeButton';
 
 // assets
 import { IconMenu2 } from '@tabler/icons-react';
@@ -37,9 +38,8 @@ const Header = ({ handleLeftDrawerToggle }) => {
             sx={{
               ...theme.typography.commonAvatar,
               ...theme.typography.mediumAvatar,
+              ...theme.typography.menuButton,
               transition: 'all .2s ease-in-out',
-              background: theme.palette.secondary.light,
-              color: theme.palette.secondary.dark,
               '&:hover': {
                 background: theme.palette.secondary.dark,
                 color: theme.palette.secondary.light
@@ -55,7 +55,7 @@ const Header = ({ handleLeftDrawerToggle }) => {
 
       <Box sx={{ flexGrow: 1 }} />
       <Box sx={{ flexGrow: 1 }} />
-
+      <ThemeButton />
       <ProfileSection />
     </>
   );
diff --git a/web/berry/src/layout/MainLayout/Sidebar/MenuCard/index.js b/web/berry/src/layout/MainLayout/Sidebar/MenuCard/index.js
index 16b13231..dadd3eca 100644
--- a/web/berry/src/layout/MainLayout/Sidebar/MenuCard/index.js
+++ b/web/berry/src/layout/MainLayout/Sidebar/MenuCard/index.js
@@ -36,7 +36,7 @@ import { useNavigate } from 'react-router-dom';
 // }));
 
 const CardStyle = styled(Card)(({ theme }) => ({
-  background: theme.palette.primary.light,
+  background: theme.typography.menuChip.background,
   marginBottom: '22px',
   overflow: 'hidden',
   position: 'relative',
@@ -121,7 +121,6 @@ const MenuCard = () => {
             />
           </ListItem>
         </List>
-        {/* <LinearProgressWithLabel value={80} /> */}
       </CardContent>
     </CardStyle>
   );
diff --git a/web/berry/src/layout/MainLayout/Sidebar/index.js b/web/berry/src/layout/MainLayout/Sidebar/index.js
index e3c6d12d..10652ba6 100644
--- a/web/berry/src/layout/MainLayout/Sidebar/index.js
+++ b/web/berry/src/layout/MainLayout/Sidebar/index.js
@@ -39,7 +39,13 @@ const Sidebar = ({ drawerOpen, drawerToggle, window }) => {
           <MenuList />
           <MenuCard />
           <Stack direction="row" justifyContent="center" sx={{ mb: 2 }}>
-            <Chip label={process.env.REACT_APP_VERSION} disabled chipcolor="secondary" size="small" sx={{ cursor: 'pointer' }} />
+            <Chip
+              label={process.env.REACT_APP_VERSION || '未知版本号'}
+              disabled
+              chipcolor="secondary"
+              size="small"
+              sx={{ cursor: 'pointer' }}
+            />
           </Stack>
         </PerfectScrollbar>
       </BrowserView>
@@ -48,7 +54,13 @@ const Sidebar = ({ drawerOpen, drawerToggle, window }) => {
           <MenuList />
           <MenuCard />
           <Stack direction="row" justifyContent="center" sx={{ mb: 2 }}>
-            <Chip label={process.env.REACT_APP_VERSION} disabled chipcolor="secondary" size="small" sx={{ cursor: 'pointer' }} />
+            <Chip
+              label={process.env.REACT_APP_VERSION || '未知版本号'}
+              disabled
+              chipcolor="secondary"
+              size="small"
+              sx={{ cursor: 'pointer' }}
+            />
           </Stack>
         </Box>
       </MobileView>
diff --git a/web/berry/src/layout/MinimalLayout/Header/index.js b/web/berry/src/layout/MinimalLayout/Header/index.js
index 4f61da60..feaeb603 100644
--- a/web/berry/src/layout/MinimalLayout/Header/index.js
+++ b/web/berry/src/layout/MinimalLayout/Header/index.js
@@ -1,10 +1,30 @@
 // material-ui
-import { useTheme } from "@mui/material/styles";
-import { Box, Button, Stack } from "@mui/material";
-import LogoSection from "layout/MainLayout/LogoSection";
-import { Link } from "react-router-dom";
-import { useLocation } from "react-router-dom";
-import { useSelector } from "react-redux";
+import { useState } from 'react';
+import { useTheme } from '@mui/material/styles';
+import {
+  Box,
+  Button,
+  Stack,
+  Popper,
+  IconButton,
+  List,
+  ListItemButton,
+  Paper,
+  ListItemText,
+  Typography,
+  Divider,
+  ClickAwayListener
+} from '@mui/material';
+import LogoSection from 'layout/MainLayout/LogoSection';
+import { Link } from 'react-router-dom';
+import { useLocation } from 'react-router-dom';
+import { useSelector } from 'react-redux';
+import ThemeButton from 'ui-component/ThemeButton';
+import ProfileSection from 'layout/MainLayout/Header/ProfileSection';
+import { IconMenu2 } from '@tabler/icons-react';
+import Transitions from 'ui-component/extended/Transitions';
+import MainCard from 'ui-component/cards/MainCard';
+import { useMediaQuery } from '@mui/material';
 
 // ==============================|| MAIN NAVBAR / HEADER ||============================== //
 
@@ -12,16 +32,26 @@ const Header = () => {
   const theme = useTheme();
   const { pathname } = useLocation();
   const account = useSelector((state) => state.account);
+  const [open, setOpen] = useState(null);
+  const isMobile = useMediaQuery(theme.breakpoints.down('sm'));
+
+  const handleOpenMenu = (event) => {
+    setOpen(open ? null : event.currentTarget);
+  };
+
+  const handleCloseMenu = () => {
+    setOpen(null);
+  };
 
   return (
     <>
       <Box
         sx={{
           width: 228,
-          display: "flex",
-          [theme.breakpoints.down("md")]: {
-            width: "auto",
-          },
+          display: 'flex',
+          [theme.breakpoints.down('md')]: {
+            width: 'auto'
+          }
         }}
       >
         <Box component="span" sx={{ flexGrow: 1 }}>
@@ -31,43 +61,99 @@ const Header = () => {
 
       <Box sx={{ flexGrow: 1 }} />
       <Box sx={{ flexGrow: 1 }} />
-      <Stack spacing={2} direction="row">
-        <Button
-          component={Link}
-          variant="text"
-          to="/"
-          color={pathname === "/" ? "primary" : "inherit"}
-        >
-          首页
-        </Button>
-        <Button
-          component={Link}
-          variant="text"
-          to="/about"
-          color={pathname === "/about" ? "primary" : "inherit"}
-        >
-          关于
-        </Button>
-        {account.user ? (
-          <Button
-            component={Link}
-            variant="contained"
-            to="/panel"
-            color="primary"
-          >
-            控制台
-          </Button>
+      <Stack spacing={2} direction="row" justifyContent="center" alignItems="center">
+        {isMobile ? (
+          <>
+            <ThemeButton />
+            <IconButton onClick={handleOpenMenu}>
+              <IconMenu2 />
+            </IconButton>
+          </>
         ) : (
-          <Button
-            component={Link}
-            variant="contained"
-            to="/login"
-            color="primary"
-          >
-            登入
-          </Button>
+          <>
+            <Button component={Link} variant="text" to="/" color={pathname === '/' ? 'primary' : 'inherit'}>
+              首页
+            </Button>
+            <Button component={Link} variant="text" to="/about" color={pathname === '/about' ? 'primary' : 'inherit'}>
+              关于
+            </Button>
+            <ThemeButton />
+            {account.user ? (
+              <>
+                <Button component={Link} variant="contained" to="/panel" color="primary">
+                  控制台
+                </Button>
+                <ProfileSection />
+              </>
+            ) : (
+              <Button component={Link} variant="contained" to="/login" color="primary">
+                登录
+              </Button>
+            )}
+          </>
         )}
       </Stack>
+
+      <Popper
+        open={!!open}
+        anchorEl={open}
+        transition
+        disablePortal
+        popperOptions={{
+          modifiers: [
+            {
+              name: 'offset',
+              options: {
+                offset: [0, 14]
+              }
+            }
+          ]
+        }}
+        style={{ width: '100vw' }}
+      >
+        {({ TransitionProps }) => (
+          <Transitions in={open} {...TransitionProps}>
+            <ClickAwayListener onClickAway={handleCloseMenu}>
+              <Paper style={{ width: '100%' }}>
+                <MainCard border={false} elevation={16} content={false} boxShadow shadow={theme.shadows[16]}>
+                  <List
+                    component="nav"
+                    sx={{
+                      width: '100%',
+                      maxWidth: '100%',
+                      minWidth: '100%',
+                      backgroundColor: theme.palette.background.paper,
+
+                      '& .MuiListItemButton-root': {
+                        mt: 0.5
+                      }
+                    }}
+                    onClick={handleCloseMenu}
+                  >
+                    <ListItemButton component={Link} variant="text" to="/">
+                      <ListItemText primary={<Typography variant="body2">首页</Typography>} />
+                    </ListItemButton>
+
+                    <ListItemButton component={Link} variant="text" to="/about">
+                      <ListItemText primary={<Typography variant="body2">关于</Typography>} />
+                    </ListItemButton>
+                    <Divider />
+                    {account.user ? (
+                      <ListItemButton component={Link} variant="contained" to="/panel" color="primary">
+                        控制台
+                      </ListItemButton>
+                    ) : (
+                      <ListItemButton component={Link} variant="contained" to="/login" color="primary">
+                        登录
+                      </ListItemButton>
+                    )}
+                  </List>
+                </MainCard>
+              </Paper>
+            </ClickAwayListener>
+          </Transitions>
+        )}
+      </Popper>
     </>
   );
 };
diff --git a/web/berry/src/layout/MinimalLayout/index.js b/web/berry/src/layout/MinimalLayout/index.js
index c2919c6d..81047fd1 100644
--- a/web/berry/src/layout/MinimalLayout/index.js
+++ b/web/berry/src/layout/MinimalLayout/index.js
@@ -1,6 +1,6 @@
 import { Outlet } from 'react-router-dom';
 import { useTheme } from '@mui/material/styles';
-import { AppBar, Box, CssBaseline, Toolbar } from '@mui/material';
+import { AppBar, Box, CssBaseline, Toolbar, Container } from '@mui/material';
 import Header from './Header';
 import Footer from 'ui-component/Footer';
 
@@ -22,9 +22,11 @@ const MinimalLayout = () => {
           flex: 'none'
         }}
       >
-        <Toolbar>
-          <Header />
-        </Toolbar>
+        <Container>
+          <Toolbar>
+            <Header />
+          </Toolbar>
+        </Container>
       </AppBar>
       <Box sx={{ flex: '1 1 auto', overflow: 'auto' }} marginTop={'80px'}>
         <Outlet />
diff --git a/web/berry/src/routes/OtherRoutes.js b/web/berry/src/routes/OtherRoutes.js
index 085c4add..a4bdb5d3 100644
--- a/web/berry/src/routes/OtherRoutes.js
+++ b/web/berry/src/routes/OtherRoutes.js
@@ -8,6 +8,8 @@ import MinimalLayout from 'layout/MinimalLayout';
 const AuthLogin = Loadable(lazy(() => import('views/Authentication/Auth/Login')));
 const AuthRegister = Loadable(lazy(() => import('views/Authentication/Auth/Register')));
 const GitHubOAuth = Loadable(lazy(() => import('views/Authentication/Auth/GitHubOAuth')));
+const LarkOAuth = Loadable(lazy(() => import('views/Authentication/Auth/LarkOAuth')));
+const OidcOAuth = Loadable(lazy(() => import('views/Authentication/Auth/OidcOAuth')));
 const ForgetPassword = Loadable(lazy(() => import('views/Authentication/Auth/ForgetPassword')));
 const ResetPassword = Loadable(lazy(() => import('views/Authentication/Auth/ResetPassword')));
 const Home = Loadable(lazy(() => import('views/Home')));
@@ -48,6 +50,14 @@ const OtherRoutes = {
       path: '/oauth/github',
       element: <GitHubOAuth />
     },
+    {
+      path: '/oauth/lark',
+      element: <LarkOAuth />
+    },
+    {
+      path: 'oauth/oidc',
+      element: <OidcOAuth />
+    },
     {
       path: '/404',
       element: <NotFoundView />
diff --git a/web/berry/src/store/actions.js b/web/berry/src/store/actions.js
index 221e8578..f1592d17 100644
--- a/web/berry/src/store/actions.js
+++ b/web/berry/src/store/actions.js
@@ -7,3 +7,4 @@ export const SET_BORDER_RADIUS = '@customization/SET_BORDER_RADIUS';
 export const SET_SITE_INFO = '@siteInfo/SET_SITE_INFO';
 export const LOGIN = '@account/LOGIN';
 export const LOGOUT = '@account/LOGOUT';
+export const SET_THEME = '@customization/SET_THEME';
diff --git a/web/berry/src/store/customizationReducer.js b/web/berry/src/store/customizationReducer.js
index bd8e5f00..0c104025 100644
--- a/web/berry/src/store/customizationReducer.js
+++ b/web/berry/src/store/customizationReducer.js
@@ -9,7 +9,8 @@ export const initialState = {
   defaultId: 'default',
   fontFamily: config.fontFamily,
   borderRadius: config.borderRadius,
-  opened: true
+  opened: true,
+  theme: 'light'
 };
 
 // ==============================|| CUSTOMIZATION REDUCER ||============================== //
@@ -38,6 +39,11 @@ const customizationReducer = (state = initialState, action) => {
         ...state,
         borderRadius: action.borderRadius
       };
+    case actionTypes.SET_THEME:
+      return {
+        ...state,
+        theme: action.theme
+      };
     default:
       return state;
   }
diff --git a/web/berry/src/themes/compStyleOverride.js b/web/berry/src/themes/compStyleOverride.js
index b6e87e01..67a3dd14 100644
--- a/web/berry/src/themes/compStyleOverride.js
+++ b/web/berry/src/themes/compStyleOverride.js
@@ -1,5 +1,5 @@
 export default function componentStyleOverrides(theme) {
-  const bgColor = theme.colors?.grey50;
+  const bgColor = theme.mode === 'dark' ? theme.backgroundDefault : theme.colors?.grey50;
   return {
     MuiButton: {
       styleOverrides: {
@@ -12,15 +12,7 @@ export default function componentStyleOverrides(theme) {
         }
       }
     },
-    MuiMenuItem: {
-      styleOverrides: {
-        root: {
-          '&:hover': {
-            backgroundColor: theme.colors?.grey100
-          }
-        }
-      }
-    }, //MuiAutocomplete-popper MuiPopover-root
+    //MuiAutocomplete-popper MuiPopover-root
     MuiAutocomplete: {
       styleOverrides: {
         popper: {
@@ -226,12 +218,12 @@ export default function componentStyleOverrides(theme) {
     MuiTableCell: {
       styleOverrides: {
         root: {
-          borderBottom: '1px solid rgb(241, 243, 244)',
+          borderBottom: '1px solid ' + theme.tableBorderBottom,
           textAlign: 'center'
         },
         head: {
           color: theme.darkTextSecondary,
-          backgroundColor: 'rgb(244, 246, 248)'
+          backgroundColor: theme.headBackgroundColor
         }
       }
     },
@@ -239,7 +231,7 @@ export default function componentStyleOverrides(theme) {
       styleOverrides: {
         root: {
           '&:hover': {
-            backgroundColor: 'rgb(244, 246, 248)'
+            backgroundColor: theme.headBackgroundColor
           }
         }
       }
@@ -247,10 +239,29 @@ export default function componentStyleOverrides(theme) {
     MuiTooltip: {
       styleOverrides: {
         tooltip: {
-          color: theme.paper,
+          color: theme.colors.paper,
           background: theme.colors?.grey700
         }
       }
+    },
+    MuiCssBaseline: {
+      styleOverrides: `
+      .apexcharts-title-text {
+          fill: ${theme.textDark} !important
+        }
+      .apexcharts-text {
+        fill: ${theme.textDark} !important
+      }
+      .apexcharts-legend-text {
+        color: ${theme.textDark} !important
+      }
+      .apexcharts-menu {
+        background: ${theme.backgroundDefault} !important
+      }
+      .apexcharts-gridline, .apexcharts-xaxistooltip-background, .apexcharts-yaxistooltip-background {
+        stroke: ${theme.divider} !important;
+      }
+      `
     }
   };
 }
diff --git a/web/berry/src/themes/index.js b/web/berry/src/themes/index.js
index 6e694aa6..addd61f7 100644
--- a/web/berry/src/themes/index.js
+++ b/web/berry/src/themes/index.js
@@ -15,19 +15,10 @@ import themeTypography from './typography';
 
 export const theme = (customization) => {
   const color = colors;
-
+  const options = customization.theme === 'light' ? GetLightOption() : GetDarkOption();
   const themeOption = {
     colors: color,
-    heading: color.grey900,
-    paper: color.paper,
-    backgroundDefault: color.paper,
-    background: color.primaryLight,
-    darkTextPrimary: color.grey700,
-    darkTextSecondary: color.grey500,
-    textDark: color.grey900,
-    menuSelected: color.secondaryDark,
-    menuSelectedBack: color.secondaryLight,
-    divider: color.grey200,
+    ...options,
     customization
   };
 
@@ -53,3 +44,49 @@ export const theme = (customization) => {
 };
 
 export default theme;
+
+function GetDarkOption() {
+  const color = colors;
+  return {
+    mode: 'dark',
+    heading: color.darkTextTitle,
+    paper: color.darkLevel2,
+    backgroundDefault: color.darkPaper,
+    background: color.darkBackground,
+    darkTextPrimary: color.darkTextPrimary,
+    darkTextSecondary: color.darkTextSecondary,
+    textDark: color.darkTextTitle,
+    menuSelected: color.darkSecondaryMain,
+    menuSelectedBack: color.darkSelectedBack,
+    divider: color.darkDivider,
+    borderColor: color.darkBorderColor,
+    menuButton: color.darkLevel1,
+    menuButtonColor: color.darkSecondaryMain,
+    menuChip: color.darkLevel1,
+    headBackgroundColor: color.darkBackground,
+    tableBorderBottom: color.darkDivider
+  };
+}
+
+function GetLightOption() {
+  const color = colors;
+  return {
+    mode: 'light',
+    heading: color.grey900,
+    paper: color.paper,
+    backgroundDefault: color.paper,
+    background: color.primaryLight,
+    darkTextPrimary: color.grey700,
+    darkTextSecondary: color.grey500,
+    textDark: color.grey900,
+    menuSelected: color.secondaryDark,
+    menuSelectedBack: color.secondaryLight,
+    divider: color.grey200,
+    borderColor: color.grey300,
+    menuButton: color.secondaryLight,
+    menuButtonColor: color.secondaryDark,
+    menuChip: color.primaryLight,
+    headBackgroundColor: color.tableBackground,
+    tableBorderBottom: color.tableBorderBottom
+  };
+}
diff --git a/web/berry/src/themes/palette.js b/web/berry/src/themes/palette.js
index 09768555..70c78782 100644
--- a/web/berry/src/themes/palette.js
+++ b/web/berry/src/themes/palette.js
@@ -5,7 +5,7 @@
 
 export default function themePalette(theme) {
   return {
-    mode: 'light',
+    mode: theme.mode,
     common: {
       black: theme.colors?.darkPaper
     },
diff --git a/web/berry/src/themes/typography.js b/web/berry/src/themes/typography.js
index 24bfabb9..f20d87a5 100644
--- a/web/berry/src/themes/typography.js
+++ b/web/berry/src/themes/typography.js
@@ -132,6 +132,19 @@ export default function themeTypography(theme) {
       width: '44px',
       height: '44px',
       fontSize: '1.5rem'
+    },
+    menuButton: {
+      color: theme.menuButtonColor,
+      background: theme.menuButton
+    },
+    menuChip: {
+      background: theme.menuChip
+    },
+    CardWrapper: {
+      backgroundColor: theme.mode === 'dark' ? theme.colors.darkLevel2 : theme.colors.primaryDark
+    },
+    SubCard: {
+      border: theme.mode === 'dark' ? '1px solid rgba(227, 232, 239, 0.2)' : '1px solid rgb(227, 232, 239)'
     }
   };
 }
diff --git a/web/berry/src/ui-component/Logo.js b/web/berry/src/ui-component/Logo.js
index a34fe895..52e61f4f 100644
--- a/web/berry/src/ui-component/Logo.js
+++ b/web/berry/src/ui-component/Logo.js
@@ -1,6 +1,8 @@
 // material-ui
-import logo from 'assets/images/logo.svg';
+import logoLight from 'assets/images/logo.svg';
+import logoDark from 'assets/images/logo-white.svg';
 import { useSelector } from 'react-redux';
+import { useTheme } from '@mui/material/styles';
 
 /**
  * if you want to use image instead of <svg> uncomment following.
@@ -14,6 +16,8 @@ import { useSelector } from 'react-redux';
 
 const Logo = () => {
   const siteInfo = useSelector((state) => state.siteInfo);
+  const theme = useTheme();
+  const logo = theme.palette.mode === 'light' ? logoLight : logoDark;
 
   return <img src={siteInfo.logo || logo} alt={siteInfo.system_name} height="50" />;
 };
diff --git a/web/berry/src/ui-component/ThemeButton.js b/web/berry/src/ui-component/ThemeButton.js
new file mode 100644
index 00000000..c907c646
--- /dev/null
+++ b/web/berry/src/ui-component/ThemeButton.js
@@ -0,0 +1,50 @@
+import { useDispatch, useSelector } from 'react-redux';
+import { SET_THEME } from 'store/actions';
+import { useTheme } from '@mui/material/styles';
+import { Avatar, Box, ButtonBase } from '@mui/material';
+import { IconSun, IconMoon } from '@tabler/icons-react';
+
+export default function ThemeButton() {
+  const dispatch = useDispatch();
+
+  const defaultTheme = useSelector((state) => state.customization.theme);
+
+  const theme = useTheme();
+
+  return (
+    <Box
+      sx={{
+        ml: 2,
+        mr: 3,
+        [theme.breakpoints.down('md')]: {
+          mr: 2
+        }
+      }}
+    >
+      <ButtonBase sx={{ borderRadius: '12px' }}>
+        <Avatar
+          variant="rounded"
+          sx={{
+            ...theme.typography.commonAvatar,
+            ...theme.typography.mediumAvatar,
+            transition: 'all .2s ease-in-out',
+            borderColor: theme.typography.menuChip.background,
+            backgroundColor: theme.typography.menuChip.background,
+            '&[aria-controls="menu-list-grow"],&:hover': {
+              background: theme.palette.secondary.dark,
+              color: theme.palette.secondary.light
+            }
+          }}
+          onClick={() => {
+            let theme = defaultTheme === 'light' ? 'dark' : 'light';
+            dispatch({ type: SET_THEME, theme: theme });
+            localStorage.setItem('theme', theme);
+          }}
+          color="inherit"
+        >
+          {defaultTheme === 'light' ? <IconSun stroke={1.5} size="1.3rem" /> : <IconMoon stroke={1.5} size="1.3rem" />}
+        </Avatar>
+      </ButtonBase>
+    </Box>
+  );
+}
diff --git a/web/berry/src/ui-component/cards/MainCard.js b/web/berry/src/ui-component/cards/MainCard.js
index 8735282c..32353027 100644
--- a/web/berry/src/ui-component/cards/MainCard.js
+++ b/web/berry/src/ui-component/cards/MainCard.js
@@ -15,7 +15,7 @@ const headerSX = {
 const MainCard = forwardRef(
   (
     {
-      border = true,
+      border = false,
       boxShadow,
       children,
       content = true,
diff --git a/web/berry/src/ui-component/cards/SubCard.js b/web/berry/src/ui-component/cards/SubCard.js
index 05f9abb7..a63819a8 100644
--- a/web/berry/src/ui-component/cards/SubCard.js
+++ b/web/berry/src/ui-component/cards/SubCard.js
@@ -15,8 +15,7 @@ const SubCard = forwardRef(
       <Card
         ref={ref}
         sx={{
-          border: '1px solid',
-          borderColor: theme.palette.primary.light,
+          border: theme.typography.SubCard.border,
           ':hover': {
             boxShadow: '0 2px 14px 0 rgb(32 40 45 / 8%)'
           },
@@ -36,8 +35,8 @@ const SubCard = forwardRef(
         {title && (
           <Divider
             sx={{
-              opacity: 1,
-              borderColor: theme.palette.primary.light
+              opacity: 1
+              // borderColor: theme.palette.primary.light
             }}
           />
         )}
@@ -62,7 +61,8 @@ SubCard.propTypes = {
   secondary: PropTypes.oneOfType([PropTypes.node, PropTypes.string, PropTypes.object]),
   sx: PropTypes.object,
   contentSX: PropTypes.object,
-  title: PropTypes.oneOfType([PropTypes.node, PropTypes.string, PropTypes.object])
+  title: PropTypes.oneOfType([PropTypes.node, PropTypes.string, PropTypes.object]),
+  subTitle: PropTypes.oneOfType([PropTypes.node, PropTypes.string, PropTypes.object])
 };
 
 SubCard.defaultProps = {
diff --git a/web/berry/src/utils/chart.js b/web/berry/src/utils/chart.js
index 4633fe37..8cf6d847 100644
--- a/web/berry/src/utils/chart.js
+++ b/web/berry/src/utils/chart.js
@@ -40,7 +40,8 @@ export function generateChartOptions(data, unit) {
       chart: {
         sparkline: {
           enabled: true
-        }
+        },
+        background: 'transparent'
       },
       dataLabels: {
         enabled: false
diff --git a/web/berry/src/utils/common.js b/web/berry/src/utils/common.js
index 25e5c635..bd85f8bf 100644
--- a/web/berry/src/utils/common.js
+++ b/web/berry/src/utils/common.js
@@ -51,9 +51,9 @@ export function showError(error) {
 
 export function showNotice(message, isHTML = false) {
   if (isHTML) {
-    enqueueSnackbar(<SnackbarHTMLContent htmlContent={message} />, getSnackbarOptions('INFO'));
+    enqueueSnackbar(<SnackbarHTMLContent htmlContent={message} />, getSnackbarOptions('NOTICE'));
   } else {
-    enqueueSnackbar(message, getSnackbarOptions('INFO'));
+    enqueueSnackbar(message, getSnackbarOptions('NOTICE'));
   }
 }
 
@@ -91,6 +91,28 @@ export async function onGitHubOAuthClicked(github_client_id, openInNewTab = fals
   }
 }
 
+export async function onLarkOAuthClicked(lark_client_id) {
+  const state = await getOAuthState();
+  if (!state) return;
+  let redirect_uri = `${window.location.origin}/oauth/lark`;
+  window.open(`https://accounts.feishu.cn/open-apis/authen/v1/authorize?redirect_uri=${redirect_uri}&client_id=${lark_client_id}&state=${state}`);
+}
+
+export async function onOidcClicked(auth_url, client_id, openInNewTab = false) {
+  const state = await getOAuthState();
+  if (!state) return;
+  const redirect_uri = `${window.location.origin}/oauth/oidc`;
+  const response_type = "code";
+  const scope = "openid profile email";
+  const url = `${auth_url}?client_id=${client_id}&redirect_uri=${redirect_uri}&response_type=${response_type}&scope=${scope}&state=${state}`;
+  if (openInNewTab) {
+    window.open(url);
+  } else
+  {
+    window.location.href = url;
+  }
+}
+
 export function isAdmin() {
   let user = localStorage.getItem('user');
   if (!user) return false;
@@ -186,3 +208,40 @@ export function removeTrailingSlash(url) {
     return url;
   }
 }
+
+let channelModels = undefined;
+export async function loadChannelModels() {
+  const res = await API.get('/api/models');
+  const { success, data } = res.data;
+  if (!success) {
+    return;
+  }
+  channelModels = data;
+  localStorage.setItem('channel_models', JSON.stringify(data));
+}
+
+export function getChannelModels(type) {
+  if (channelModels !== undefined && type in channelModels) {
+    return channelModels[type];
+  }
+  let models = localStorage.getItem('channel_models');
+  if (!models) {
+    return [];
+  }
+  channelModels = JSON.parse(models);
+  if (type in channelModels) {
+    return channelModels[type];
+  }
+  return [];
+}
+
+export function copy(text, name = '') {
+  try {
+    navigator.clipboard.writeText(text);
+  } catch (error) {
+    text = `复制${name}失败，请手动复制：<br /><br />${text}`;
+    enqueueSnackbar(<SnackbarHTMLContent htmlContent={text} />, getSnackbarOptions('COPY'));
+    return;
+  }
+  showSuccess(`复制${name}成功！`);
+}
diff --git a/web/berry/src/views/Authentication/Auth/LarkOAuth.js b/web/berry/src/views/Authentication/Auth/LarkOAuth.js
new file mode 100644
index 00000000..88ced5d8
--- /dev/null
+++ b/web/berry/src/views/Authentication/Auth/LarkOAuth.js
@@ -0,0 +1,94 @@
+import { Link, useNavigate, useSearchParams } from 'react-router-dom';
+import React, { useEffect, useState } from 'react';
+import { showError } from 'utils/common';
+import useLogin from 'hooks/useLogin';
+
+// material-ui
+import { useTheme } from '@mui/material/styles';
+import { Grid, Stack, Typography, useMediaQuery, CircularProgress } from '@mui/material';
+
+// project imports
+import AuthWrapper from '../AuthWrapper';
+import AuthCardWrapper from '../AuthCardWrapper';
+import Logo from 'ui-component/Logo';
+
+// assets
+
+// ================================|| AUTH3 - LOGIN ||================================ //
+
+const LarkOAuth = () => {
+  const theme = useTheme();
+  const matchDownSM = useMediaQuery(theme.breakpoints.down('md'));
+
+  const [searchParams] = useSearchParams();
+  const [prompt, setPrompt] = useState('处理中...');
+  const { larkLogin } = useLogin();
+
+  let navigate = useNavigate();
+
+  const sendCode = async (code, state, count) => {
+    const { success, message } = await larkLogin(code, state);
+    if (!success) {
+      if (message) {
+        showError(message);
+      }
+      if (count === 0) {
+        setPrompt(`操作失败，重定向至登录界面中...`);
+        await new Promise((resolve) => setTimeout(resolve, 2000));
+        navigate('/login');
+        return;
+      }
+      count++;
+      setPrompt(`出现错误，第 ${count} 次重试中...`);
+      await new Promise((resolve) => setTimeout(resolve, 2000));
+      await sendCode(code, state, count);
+    }
+  };
+
+  useEffect(() => {
+    let code = searchParams.get('code');
+    let state = searchParams.get('state');
+    sendCode(code, state, 0).then();
+  }, []);
+
+  return (
+    <AuthWrapper>
+      <Grid container direction="column" justifyContent="flex-end">
+        <Grid item xs={12}>
+          <Grid container justifyContent="center" alignItems="center" sx={{ minHeight: 'calc(100vh - 136px)' }}>
+            <Grid item sx={{ m: { xs: 1, sm: 3 }, mb: 0 }}>
+              <AuthCardWrapper>
+                <Grid container spacing={2} alignItems="center" justifyContent="center">
+                  <Grid item sx={{ mb: 3 }}>
+                    <Link to="#">
+                      <Logo />
+                    </Link>
+                  </Grid>
+                  <Grid item xs={12}>
+                    <Grid container direction={matchDownSM ? 'column-reverse' : 'row'} alignItems="center" justifyContent="center">
+                      <Grid item>
+                        <Stack alignItems="center" justifyContent="center" spacing={1}>
+                          <Typography color={theme.palette.primary.main} gutterBottom variant={matchDownSM ? 'h3' : 'h2'}>
+                            飞书 登录
+                          </Typography>
+                        </Stack>
+                      </Grid>
+                    </Grid>
+                  </Grid>
+                  <Grid item xs={12} container direction="column" justifyContent="center" alignItems="center" style={{ height: '200px' }}>
+                    <CircularProgress />
+                    <Typography variant="h3" paddingTop={'20px'}>
+                      {prompt}
+                    </Typography>
+                  </Grid>
+                </Grid>
+              </AuthCardWrapper>
+            </Grid>
+          </Grid>
+        </Grid>
+      </Grid>
+    </AuthWrapper>
+  );
+};
+
+export default LarkOAuth;
diff --git a/web/berry/src/views/Authentication/Auth/OidcOAuth.js b/web/berry/src/views/Authentication/Auth/OidcOAuth.js
new file mode 100644
index 00000000..55d9372d
--- /dev/null
+++ b/web/berry/src/views/Authentication/Auth/OidcOAuth.js
@@ -0,0 +1,94 @@
+import { Link, useNavigate, useSearchParams } from 'react-router-dom';
+import React, { useEffect, useState } from 'react';
+import { showError } from 'utils/common';
+import useLogin from 'hooks/useLogin';
+
+// material-ui
+import { useTheme } from '@mui/material/styles';
+import { Grid, Stack, Typography, useMediaQuery, CircularProgress } from '@mui/material';
+
+// project imports
+import AuthWrapper from '../AuthWrapper';
+import AuthCardWrapper from '../AuthCardWrapper';
+import Logo from 'ui-component/Logo';
+
+// assets
+
+// ================================|| AUTH3 - LOGIN ||================================ //
+
+const OidcOAuth = () => {
+  const theme = useTheme();
+  const matchDownSM = useMediaQuery(theme.breakpoints.down('md'));
+
+  const [searchParams] = useSearchParams();
+  const [prompt, setPrompt] = useState('处理中...');
+  const { oidcLogin } = useLogin();
+
+  let navigate = useNavigate();
+
+  const sendCode = async (code, state, count) => {
+    const { success, message } = await oidcLogin(code, state);
+    if (!success) {
+      if (message) {
+        showError(message);
+      }
+      if (count === 0) {
+        setPrompt(`操作失败，重定向至登录界面中...`);
+        await new Promise((resolve) => setTimeout(resolve, 2000));
+        navigate('/login');
+        return;
+      }
+      count++;
+      setPrompt(`出现错误，第 ${count} 次重试中...`);
+      await new Promise((resolve) => setTimeout(resolve, 2000));
+      await sendCode(code, state, count);
+    }
+  };
+
+  useEffect(() => {
+    let code = searchParams.get('code');
+    let state = searchParams.get('state');
+    sendCode(code, state, 0).then();
+  }, []);
+
+  return (
+    <AuthWrapper>
+      <Grid container direction="column" justifyContent="flex-end">
+        <Grid item xs={12}>
+          <Grid container justifyContent="center" alignItems="center" sx={{ minHeight: 'calc(100vh - 136px)' }}>
+            <Grid item sx={{ m: { xs: 1, sm: 3 }, mb: 0 }}>
+              <AuthCardWrapper>
+                <Grid container spacing={2} alignItems="center" justifyContent="center">
+                  <Grid item sx={{ mb: 3 }}>
+                    <Link to="#">
+                      <Logo />
+                    </Link>
+                  </Grid>
+                  <Grid item xs={12}>
+                    <Grid container direction={matchDownSM ? 'column-reverse' : 'row'} alignItems="center" justifyContent="center">
+                      <Grid item>
+                        <Stack alignItems="center" justifyContent="center" spacing={1}>
+                          <Typography color={theme.palette.primary.main} gutterBottom variant={matchDownSM ? 'h3' : 'h2'}>
+                            OIDC 登录
+                          </Typography>
+                        </Stack>
+                      </Grid>
+                    </Grid>
+                  </Grid>
+                  <Grid item xs={12} container direction="column" justifyContent="center" alignItems="center" style={{ height: '200px' }}>
+                    <CircularProgress />
+                    <Typography variant="h3" paddingTop={'20px'}>
+                      {prompt}
+                    </Typography>
+                  </Grid>
+                </Grid>
+              </AuthCardWrapper>
+            </Grid>
+          </Grid>
+        </Grid>
+      </Grid>
+    </AuthWrapper>
+  );
+};
+
+export default OidcOAuth;
diff --git a/web/berry/src/views/Authentication/Auth/Register.js b/web/berry/src/views/Authentication/Auth/Register.js
index 4489e560..8027649d 100644
--- a/web/berry/src/views/Authentication/Auth/Register.js
+++ b/web/berry/src/views/Authentication/Auth/Register.js
@@ -51,7 +51,7 @@ const Register = () => {
                   <Grid item xs={12}>
                     <Grid item container direction="column" alignItems="center" xs={12}>
                       <Typography component={Link} to="/login" variant="subtitle1" sx={{ textDecoration: 'none' }}>
-                        已经有帐号了?点击登录
+                        已经有帐号了？点击登录
                       </Typography>
                     </Grid>
                   </Grid>
diff --git a/web/berry/src/views/Authentication/AuthForms/AuthLogin.js b/web/berry/src/views/Authentication/AuthForms/AuthLogin.js
index 9420b098..7efd0362 100644
--- a/web/berry/src/views/Authentication/AuthForms/AuthLogin.js
+++ b/web/berry/src/views/Authentication/AuthForms/AuthLogin.js
@@ -35,7 +35,9 @@ import VisibilityOff from '@mui/icons-material/VisibilityOff';
 
 import Github from 'assets/images/icons/github.svg';
 import Wechat from 'assets/images/icons/wechat.svg';
-import { onGitHubOAuthClicked } from 'utils/common';
+import Lark from 'assets/images/icons/lark.svg';
+import OIDC from 'assets/images/icons/oidc.svg';
+import { onGitHubOAuthClicked, onLarkOAuthClicked, onOidcClicked } from 'utils/common';
 
 // ============================|| FIREBASE - LOGIN ||============================ //
 
@@ -49,7 +51,7 @@ const LoginForm = ({ ...others }) => {
   // const [checked, setChecked] = useState(true);
 
   let tripartiteLogin = false;
-  if (siteInfo.github_oauth || siteInfo.wechat_login) {
+  if (siteInfo.github_oauth || siteInfo.wechat_login || siteInfo.lark_client_id || siteInfo.oidc) {
     tripartiteLogin = true;
   }
 
@@ -121,6 +123,52 @@ const LoginForm = ({ ...others }) => {
               <WechatModal open={openWechat} handleClose={handleWechatClose} wechatLogin={wechatLogin} qrCode={siteInfo.wechat_qrcode} />
             </Grid>
           )}
+          {siteInfo.lark_client_id && (
+            <Grid item xs={12}>
+              <AnimateButton>
+                <Button
+                  disableElevation
+                  fullWidth
+                  onClick={() => onLarkOAuthClicked(siteInfo.lark_client_id)}
+                  size="large"
+                  variant="outlined"
+                  sx={{
+                    color: 'grey.700',
+                    backgroundColor: theme.palette.grey[50],
+                    borderColor: theme.palette.grey[100]
+                  }}
+                >
+                  <Box sx={{ mr: { xs: 1, sm: 2, width: 20 }, display: 'flex', alignItems: 'center' }}>
+                    <img src={Lark} alt="Lark" width={25} height={25} style={{ marginRight: matchDownSM ? 8 : 16 }} />
+                  </Box>
+                  使用飞书登录
+                </Button>
+              </AnimateButton>
+            </Grid>
+          )}
+          {siteInfo.oidc && (
+            <Grid item xs={12}>
+              <AnimateButton>
+                <Button
+                  disableElevation
+                  fullWidth
+                  onClick={() => onOidcClicked(siteInfo.oidc_authorization_endpoint,siteInfo.oidc_client_id)}
+                  size="large"
+                  variant="outlined"
+                  sx={{
+                    color: 'grey.700',
+                    backgroundColor: theme.palette.grey[50],
+                    borderColor: theme.palette.grey[100]
+                  }}
+                >
+                  <Box sx={{ mr: { xs: 1, sm: 2, width: 20 }, display: 'flex', alignItems: 'center' }}>
+                    <img src={OIDC} alt="Lark" width={25} height={25} style={{ marginRight: matchDownSM ? 8 : 16 }} />
+                  </Box>
+                  使用 OIDC 登录
+                </Button>
+              </AnimateButton>
+            </Grid>
+          )}
           <Grid item xs={12}>
             <Box
               sx={{
diff --git a/web/berry/src/views/Authentication/AuthForms/AuthRegister.js b/web/berry/src/views/Authentication/AuthForms/AuthRegister.js
index c286faad..8d588696 100644
--- a/web/berry/src/views/Authentication/AuthForms/AuthRegister.js
+++ b/web/berry/src/views/Authentication/AuthForms/AuthRegister.js
@@ -296,7 +296,7 @@ const RegisterForm = ({ ...others }) => {
             <Box sx={{ mt: 2 }}>
               <AnimateButton>
                 <Button disableElevation disabled={isSubmitting} fullWidth size="large" type="submit" variant="contained" color="primary">
-                  Sign up
+                  注册
                 </Button>
               </AnimateButton>
             </Box>
diff --git a/web/berry/src/views/Authentication/AuthForms/ResetPasswordForm.js b/web/berry/src/views/Authentication/AuthForms/ResetPasswordForm.js
index eaa8dc95..a9f0f9e3 100644
--- a/web/berry/src/views/Authentication/AuthForms/ResetPasswordForm.js
+++ b/web/berry/src/views/Authentication/AuthForms/ResetPasswordForm.js
@@ -1,22 +1,22 @@
-import { useState, useEffect } from "react";
-import { useSearchParams } from "react-router-dom";
+import { useState, useEffect } from 'react';
+import { useSearchParams } from 'react-router-dom';
 
 // material-ui
-import { Button, Stack, Typography, Alert } from "@mui/material";
+import { Button, Stack, Typography, Alert } from '@mui/material';
 
 // assets
-import { showError, showInfo } from "utils/common";
-import { API } from "utils/api";
+import { showError, copy } from 'utils/common';
+import { API } from 'utils/api';
 
 // ===========================|| FIREBASE - REGISTER ||=========================== //
 
 const ResetPasswordForm = () => {
   const [searchParams] = useSearchParams();
   const [inputs, setInputs] = useState({
-    email: "",
-    token: "",
+    email: '',
+    token: ''
   });
-  const [newPassword, setNewPassword] = useState("");
+  const [newPassword, setNewPassword] = useState('');
 
   const submit = async () => {
     const res = await API.post(`/api/user/reset`, inputs);
@@ -24,31 +24,25 @@ const ResetPasswordForm = () => {
     if (success) {
       let password = res.data.data;
       setNewPassword(password);
-      navigator.clipboard.writeText(password);
-      showInfo(`新密码已复制到剪贴板：${password}`);
+      copy(password, '新密码');
     } else {
       showError(message);
     }
   };
 
   useEffect(() => {
-    let email = searchParams.get("email");
-    let token = searchParams.get("token");
+    let email = searchParams.get('email');
+    let token = searchParams.get('token');
     setInputs({
       token,
-      email,
+      email
     });
   }, []);
 
   return (
-    <Stack
-      spacing={3}
-      padding={"24px"}
-      justifyContent={"center"}
-      alignItems={"center"}
-    >
+    <Stack spacing={3} padding={'24px'} justifyContent={'center'} alignItems={'center'}>
       {!inputs.email || !inputs.token ? (
-        <Typography variant="h3" sx={{ textDecoration: "none" }}>
+        <Typography variant="h3" sx={{ textDecoration: 'none' }}>
           无效的链接
         </Typography>
       ) : newPassword ? (
@@ -57,14 +51,7 @@ const ResetPasswordForm = () => {
           请登录后及时修改密码
         </Alert>
       ) : (
-        <Button
-          fullWidth
-          onClick={submit}
-          size="large"
-          type="submit"
-          variant="contained"
-          color="primary"
-        >
+        <Button fullWidth onClick={submit} size="large" type="submit" variant="contained" color="primary">
           点击重置密码
         </Button>
       )}
diff --git a/web/berry/src/views/Authentication/AuthWrapper.js b/web/berry/src/views/Authentication/AuthWrapper.js
index 8cd0ec29..dc875704 100644
--- a/web/berry/src/views/Authentication/AuthWrapper.js
+++ b/web/berry/src/views/Authentication/AuthWrapper.js
@@ -8,7 +8,7 @@ import { UserContext } from 'contexts/UserContext';
 // ==============================|| AUTHENTICATION 1 WRAPPER ||============================== //
 
 const AuthStyle = styled('div')(({ theme }) => ({
-  backgroundColor: theme.palette.primary.light
+  backgroundColor: theme.palette.background.default
 }));
 
 // eslint-disable-next-line
diff --git a/web/berry/src/views/Channel/component/EditModal.js b/web/berry/src/views/Channel/component/EditModal.js
index 07111c97..77f7219a 100644
--- a/web/berry/src/views/Channel/component/EditModal.js
+++ b/web/berry/src/views/Channel/component/EditModal.js
@@ -1,9 +1,9 @@
-import PropTypes from "prop-types";
-import { useState, useEffect } from "react";
-import { CHANNEL_OPTIONS } from "constants/ChannelConstants";
-import { useTheme } from "@mui/material/styles";
-import { API } from "utils/api";
-import { showError, showSuccess } from "utils/common";
+import PropTypes from 'prop-types';
+import { useState, useEffect } from 'react';
+import { CHANNEL_OPTIONS } from 'constants/ChannelConstants';
+import { useTheme } from '@mui/material/styles';
+import { API } from 'utils/api';
+import { showError, showSuccess, getChannelModels } from 'utils/common';
 import {
   Dialog,
   DialogTitle,
@@ -21,13 +21,14 @@ import {
   Container,
   Autocomplete,
   FormHelperText,
+  Switch,
   Checkbox
-} from "@mui/material";
+} from '@mui/material';
 
-import { Formik } from "formik";
-import * as Yup from "yup";
-import { defaultConfig, typeConfig } from "../type/Config"; //typeConfig
-import { createFilterOptions } from "@mui/material/Autocomplete";
+import { Formik } from 'formik';
+import * as Yup from 'yup';
+import { defaultConfig, typeConfig } from '../type/Config'; //typeConfig
+import { createFilterOptions } from '@mui/material/Autocomplete';
 import CheckBoxOutlineBlankIcon from '@mui/icons-material/CheckBoxOutlineBlank';
 import CheckBoxIcon from '@mui/icons-material/CheckBox';
 
@@ -37,38 +38,34 @@ const checkedIcon = <CheckBoxIcon fontSize="small" />;
 const filter = createFilterOptions();
 const validationSchema = Yup.object().shape({
   is_edit: Yup.boolean(),
-  name: Yup.string().required("名称 不能为空"),
-  type: Yup.number().required("渠道 不能为空"),
-  key: Yup.string().when("is_edit", {
-    is: false,
-    then: Yup.string().required("密钥 不能为空"),
+  name: Yup.string().required('名称 不能为空'),
+  type: Yup.number().required('渠道 不能为空'),
+  key: Yup.string().when(['is_edit', 'type'], {
+    is: (is_edit, type) => !is_edit && type !== 33,
+    then: Yup.string().required('密钥 不能为空')
   }),
   other: Yup.string(),
-  models: Yup.array().min(1, "模型 不能为空"),
-  groups: Yup.array().min(1, "用户组 不能为空"),
-  base_url: Yup.string().when("type", {
+  models: Yup.array().min(1, '模型 不能为空'),
+  groups: Yup.array().min(1, '用户组 不能为空'),
+  base_url: Yup.string().when('type', {
     is: (value) => [3, 8].includes(value),
-    then: Yup.string().required("渠道API地址 不能为空"), // base_url 是必需的
-    otherwise: Yup.string(), // 在其他情况下，base_url 可以是任意字符串
+    then: Yup.string().required('渠道API地址 不能为空'), // base_url 是必需的
+    otherwise: Yup.string() // 在其他情况下，base_url 可以是任意字符串
   }),
-  model_mapping: Yup.string().test(
-    "is-json",
-    "必须是有效的JSON字符串",
-    function (value) {
-      try {
-        if (value === "" || value === null || value === undefined) {
-          return true;
-        }
-        const parsedValue = JSON.parse(value);
-        if (typeof parsedValue === "object") {
-          return true;
-        }
-      } catch (e) {
-        return false;
+  model_mapping: Yup.string().test('is-json', '必须是有效的JSON字符串', function (value) {
+    try {
+      if (value === '' || value === null || value === undefined) {
+        return true;
       }
+      const parsedValue = JSON.parse(value);
+      if (typeof parsedValue === 'object') {
+        return true;
+      }
+    } catch (e) {
       return false;
     }
-  ),
+    return false;
+  })
 });
 
 const EditModal = ({ open, channelId, onCancel, onOk }) => {
@@ -79,12 +76,14 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
   const [inputPrompt, setInputPrompt] = useState(defaultConfig.prompt);
   const [groupOptions, setGroupOptions] = useState([]);
   const [modelOptions, setModelOptions] = useState([]);
+  const [batchAdd, setBatchAdd] = useState(false);
+  const [basicModels, setBasicModels] = useState([]);
 
   const initChannel = (typeValue) => {
     if (typeConfig[typeValue]?.inputLabel) {
       setInputLabel({
         ...defaultConfig.inputLabel,
-        ...typeConfig[typeValue].inputLabel,
+        ...typeConfig[typeValue].inputLabel
       });
     } else {
       setInputLabel(defaultConfig.inputLabel);
@@ -93,7 +92,7 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     if (typeConfig[typeValue]?.prompt) {
       setInputPrompt({
         ...defaultConfig.prompt,
-        ...typeConfig[typeValue].prompt,
+        ...typeConfig[typeValue].prompt
       });
     } else {
       setInputPrompt(defaultConfig.prompt);
@@ -102,40 +101,14 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     return typeConfig[typeValue]?.input;
   };
   const handleTypeChange = (setFieldValue, typeValue, values) => {
-    const newInput = initChannel(typeValue);
-
-    if (newInput) {
-      Object.keys(newInput).forEach((key) => {
-        if (
-          (!Array.isArray(values[key]) &&
-            values[key] !== null &&
-            values[key] !== undefined &&
-            values[key] !== "") ||
-          (Array.isArray(values[key]) && values[key].length > 0)
-        ) {
-          return;
-        }
-
-        if (key === "models") {
-          setFieldValue(key, initialModel(newInput[key]));
-          return;
-        }
-        setFieldValue(key, newInput[key]);
-      });
+    initChannel(typeValue);
+    let localModels = getChannelModels(typeValue);
+    setBasicModels(localModels);
+    if (localModels.length > 0 && Array.isArray(values['models']) && values['models'].length == 0) {
+      setFieldValue('models', initialModel(localModels));
     }
-  };
 
-  const basicModels = (channelType) => {
-    let modelGroup =
-      typeConfig[channelType]?.modelGroup || defaultConfig.modelGroup;
-    // 循环 modelOptions，找到 modelGroup 对应的模型
-    let modelList = [];
-    modelOptions.forEach((model) => {
-      if (model.group === modelGroup) {
-        modelList.push(model);
-      }
-    });
-    return modelList;
+    setFieldValue('config', {});
   };
 
   const fetchGroups = async () => {
@@ -151,9 +124,9 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     try {
       let res = await API.get(`/api/channel/models`);
       const { data } = res.data;
-      data.forEach(item => {
+      data.forEach((item) => {
         if (!item.owned_by) {
-          item.owned_by = "未知";
+          item.owned_by = '未知';
         }
       });
       // 先对data排序
@@ -166,10 +139,10 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
       });
 
       setModelOptions(
-         data.map((model) => {
+        data.map((model) => {
           return {
             id: model.id,
-            group: model.owned_by,
+            group: model.owned_by
           };
         })
       );
@@ -180,33 +153,43 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
 
   const submit = async (values, { setErrors, setStatus, setSubmitting }) => {
     setSubmitting(true);
-    if (values.base_url && values.base_url.endsWith("/")) {
+    if (values.base_url && values.base_url.endsWith('/')) {
       values.base_url = values.base_url.slice(0, values.base_url.length - 1);
     }
-    if (values.type === 3 && values.other === "") {
-      values.other = "2023-09-01-preview";
+    if (values.type === 3 && values.other === '') {
+      values.other = '2023-09-01-preview';
     }
-    if (values.type === 18 && values.other === "") {
-      values.other = "v2.1";
+    if (values.type === 18 && values.other === '') {
+      values.other = 'v2.1';
     }
+    if (values.key === '') {
+      if (values.config.ak && values.config.sk && values.config.region) {
+        values.key = `${values.config.ak}|${values.config.sk}|${values.config.region}`;
+      } else if (values.config.region && values.config.vertex_ai_project_id && values.config.vertex_ai_adc) {
+        values.key = `${values.config.region}|${values.config.vertex_ai_project_id}|${values.config.vertex_ai_adc}`;
+      }
+    }
+
     let res;
-    const modelsStr = values.models.map((model) => model.id).join(",");
-    values.group = values.groups.join(",");
+    const modelsStr = values.models.map((model) => model.id).join(',');
+    const configStr = JSON.stringify(values.config);
+    values.group = values.groups.join(',');
     if (channelId) {
       res = await API.put(`/api/channel/`, {
         ...values,
         id: parseInt(channelId),
         models: modelsStr,
+        config: configStr
       });
     } else {
-      res = await API.post(`/api/channel/`, { ...values, models: modelsStr });
+      res = await API.post(`/api/channel/`, { ...values, models: modelsStr, config: configStr });
     }
     const { success, message } = res.data;
     if (success) {
       if (channelId) {
-        showSuccess("渠道更新成功！");
+        showSuccess('渠道更新成功！');
       } else {
-        showSuccess("渠道创建成功！");
+        showSuccess('渠道创建成功！');
       }
       setSubmitting(false);
       setStatus({ success: true });
@@ -224,15 +207,15 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     }
 
     // 如果 channelModel 是一个字符串
-    if (typeof channelModel === "string") {
-      channelModel = channelModel.split(",");
+    if (typeof channelModel === 'string') {
+      channelModel = channelModel.split(',');
     }
     let modelList = channelModel.map((model) => {
       const modelOption = modelOptions.find((option) => option.id === model);
       if (modelOption) {
         return modelOption;
       }
-      return { id: model, group: "自定义：点击或回车输入" };
+      return { id: model, group: '自定义：点击或回车输入' };
     });
     return modelList;
   }
@@ -241,23 +224,23 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
     let res = await API.get(`/api/channel/${channelId}`);
     const { success, message, data } = res.data;
     if (success) {
-      if (data.models === "") {
+      if (data.models === '') {
         data.models = [];
       } else {
         data.models = initialModel(data.models);
       }
-      if (data.group === "") {
+      if (data.group === '') {
         data.groups = [];
       } else {
-        data.groups = data.group.split(",");
+        data.groups = data.group.split(',');
       }
-      if (data.model_mapping !== "") {
-        data.model_mapping = JSON.stringify(
-          JSON.parse(data.model_mapping),
-          null,
-          2
-        );
+      if (data.model_mapping !== '') {
+        data.model_mapping = JSON.stringify(JSON.parse(data.model_mapping), null, 2);
       }
+      if (data.config !== '') {
+        data.config = JSON.parse(data.config);
+      }
+
       data.base_url = data.base_url ?? '';
       data.is_edit = true;
       initChannel(data.type);
@@ -273,6 +256,7 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
   }, []);
 
   useEffect(() => {
+    setBatchAdd(false);
     if (channelId) {
       loadChannel().then();
     } else {
@@ -283,45 +267,25 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
   }, [channelId]);
 
   return (
-    <Dialog open={open} onClose={onCancel} fullWidth maxWidth={"md"}>
+    <Dialog open={open} onClose={onCancel} fullWidth maxWidth={'md'}>
       <DialogTitle
         sx={{
-          margin: "0px",
+          margin: '0px',
           fontWeight: 700,
-          lineHeight: "1.55556",
-          padding: "24px",
-          fontSize: "1.125rem",
+          lineHeight: '1.55556',
+          padding: '24px',
+          fontSize: '1.125rem'
         }}
       >
-        {channelId ? "编辑渠道" : "新建渠道"}
+        {channelId ? '编辑渠道' : '新建渠道'}
       </DialogTitle>
       <Divider />
       <DialogContent>
-        <Formik
-          initialValues={initialInput}
-          enableReinitialize
-          validationSchema={validationSchema}
-          onSubmit={submit}
-        >
-          {({
-            errors,
-            handleBlur,
-            handleChange,
-            handleSubmit,
-            isSubmitting,
-            touched,
-            values,
-            setFieldValue,
-          }) => (
+        <Formik initialValues={initialInput} enableReinitialize validationSchema={validationSchema} onSubmit={submit}>
+          {({ errors, handleBlur, handleChange, handleSubmit, isSubmitting, touched, values, setFieldValue }) => (
             <form noValidate onSubmit={handleSubmit}>
-              <FormControl
-                fullWidth
-                error={Boolean(touched.type && errors.type)}
-                sx={{ ...theme.typography.otherInput }}
-              >
-                <InputLabel htmlFor="channel-type-label">
-                  {inputLabel.type}
-                </InputLabel>
+              <FormControl fullWidth error={Boolean(touched.type && errors.type)} sx={{ ...theme.typography.otherInput }}>
+                <InputLabel htmlFor="channel-type-label">{inputLabel.type}</InputLabel>
                 <Select
                   id="channel-type-label"
                   label={inputLabel.type}
@@ -335,39 +299,34 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   MenuProps={{
                     PaperProps: {
                       style: {
-                        maxHeight: 200,
-                      },
-                    },
+                        maxHeight: 200
+                      }
+                    }
                   }}
                 >
-                  {Object.values(CHANNEL_OPTIONS).map((option) => {
-                    return (
-                      <MenuItem key={option.value} value={option.value}>
-                        {option.text}
-                      </MenuItem>
-                    );
-                  })}
+                  {Object.values(CHANNEL_OPTIONS)
+                    .sort((a, b) => {
+                      return a.text.localeCompare(b.text);
+                    })
+                    .map((option) => {
+                      return (
+                        <MenuItem key={option.value} value={option.value}>
+                          {option.text}
+                        </MenuItem>
+                      );
+                    })}
                 </Select>
                 {touched.type && errors.type ? (
                   <FormHelperText error id="helper-tex-channel-type-label">
                     {errors.type}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-type-label">
-                    {" "}
-                    {inputPrompt.type}{" "}
-                  </FormHelperText>
+                  <FormHelperText id="helper-tex-channel-type-label"> {inputPrompt.type} </FormHelperText>
                 )}
               </FormControl>
 
-              <FormControl
-                fullWidth
-                error={Boolean(touched.name && errors.name)}
-                sx={{ ...theme.typography.otherInput }}
-              >
-                <InputLabel htmlFor="channel-name-label">
-                  {inputLabel.name}
-                </InputLabel>
+              <FormControl fullWidth error={Boolean(touched.name && errors.name)} sx={{ ...theme.typography.otherInput }}>
+                <InputLabel htmlFor="channel-name-label">{inputLabel.name}</InputLabel>
                 <OutlinedInput
                   id="channel-name-label"
                   label={inputLabel.name}
@@ -376,7 +335,7 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   name="name"
                   onBlur={handleBlur}
                   onChange={handleChange}
-                  inputProps={{ autoComplete: "name" }}
+                  inputProps={{ autoComplete: 'name' }}
                   aria-describedby="helper-text-channel-name-label"
                 />
                 {touched.name && errors.name ? (
@@ -384,21 +343,12 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                     {errors.name}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-name-label">
-                    {" "}
-                    {inputPrompt.name}{" "}
-                  </FormHelperText>
+                  <FormHelperText id="helper-tex-channel-name-label"> {inputPrompt.name} </FormHelperText>
                 )}
               </FormControl>
 
-              <FormControl
-                fullWidth
-                error={Boolean(touched.base_url && errors.base_url)}
-                sx={{ ...theme.typography.otherInput }}
-              >
-                <InputLabel htmlFor="channel-base_url-label">
-                  {inputLabel.base_url}
-                </InputLabel>
+              <FormControl fullWidth error={Boolean(touched.base_url && errors.base_url)} sx={{ ...theme.typography.otherInput }}>
+                <InputLabel htmlFor="channel-base_url-label">{inputLabel.base_url}</InputLabel>
                 <OutlinedInput
                   id="channel-base_url-label"
                   label={inputLabel.base_url}
@@ -415,22 +365,13 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                     {errors.base_url}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-base_url-label">
-                    {" "}
-                    {inputPrompt.base_url}{" "}
-                  </FormHelperText>
+                  <FormHelperText id="helper-tex-channel-base_url-label"> {inputPrompt.base_url} </FormHelperText>
                 )}
               </FormControl>
 
               {inputPrompt.other && (
-                <FormControl
-                  fullWidth
-                  error={Boolean(touched.other && errors.other)}
-                  sx={{ ...theme.typography.otherInput }}
-                >
-                  <InputLabel htmlFor="channel-other-label">
-                    {inputLabel.other}
-                  </InputLabel>
+                <FormControl fullWidth error={Boolean(touched.other && errors.other)} sx={{ ...theme.typography.otherInput }}>
+                  <InputLabel htmlFor="channel-other-label">{inputLabel.other}</InputLabel>
                   <OutlinedInput
                     id="channel-other-label"
                     label={inputLabel.other}
@@ -447,10 +388,7 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                       {errors.other}
                     </FormHelperText>
                   ) : (
-                    <FormHelperText id="helper-tex-channel-other-label">
-                      {" "}
-                      {inputPrompt.other}{" "}
-                    </FormHelperText>
+                    <FormHelperText id="helper-tex-channel-other-label"> {inputPrompt.other} </FormHelperText>
                   )}
                 </FormControl>
               )}
@@ -464,22 +402,15 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   onChange={(e, value) => {
                     const event = {
                       target: {
-                        name: "groups",
-                        value: value,
-                      },
+                        name: 'groups',
+                        value: value
+                      }
                     };
                     handleChange(event);
                   }}
                   onBlur={handleBlur}
                   filterSelectedOptions
-                  renderInput={(params) => (
-                    <TextField
-                      {...params}
-                      name="groups"
-                      error={Boolean(errors.groups)}
-                      label={inputLabel.groups}
-                    />
-                  )}
+                  renderInput={(params) => <TextField {...params} name="groups" error={Boolean(errors.groups)} label={inputLabel.groups} />}
                   aria-describedby="helper-text-channel-groups-label"
                 />
                 {errors.groups ? (
@@ -487,10 +418,7 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                     {errors.groups}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-groups-label">
-                    {" "}
-                    {inputPrompt.groups}{" "}
-                  </FormHelperText>
+                  <FormHelperText id="helper-tex-channel-groups-label"> {inputPrompt.groups} </FormHelperText>
                 )}
               </FormControl>
 
@@ -504,30 +432,19 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   onChange={(e, value) => {
                     const event = {
                       target: {
-                        name: "models",
-                        value: value.map((item) =>
-                          typeof item === "string"
-                            ? { id: item, group: "自定义：点击或回车输入" }
-                            : item
-                        ),
-                      },
+                        name: 'models',
+                        value: value.map((item) => (typeof item === 'string' ? { id: item, group: '自定义：点击或回车输入' } : item))
+                      }
                     };
                     handleChange(event);
                   }}
                   onBlur={handleBlur}
                   // filterSelectedOptions
                   disableCloseOnSelect
-                  renderInput={(params) => (
-                    <TextField
-                      {...params}
-                      name="models"
-                      error={Boolean(errors.models)}
-                      label={inputLabel.models}
-                    />
-                  )}
+                  renderInput={(params) => <TextField {...params} name="models" error={Boolean(errors.models)} label={inputLabel.models} />}
                   groupBy={(option) => option.group}
                   getOptionLabel={(option) => {
-                    if (typeof option === "string") {
+                    if (typeof option === 'string') {
                       return option;
                     }
                     if (option.inputValue) {
@@ -538,13 +455,11 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   filterOptions={(options, params) => {
                     const filtered = filter(options, params);
                     const { inputValue } = params;
-                    const isExisting = options.some(
-                      (option) => inputValue === option.id
-                    );
-                    if (inputValue !== "" && !isExisting) {
+                    const isExisting = options.some((option) => inputValue === option.id);
+                    if (inputValue !== '' && !isExisting) {
                       filtered.push({
                         id: inputValue,
-                        group: "自定义：点击或回车输入",
+                        group: '自定义：点击或回车输入'
                       });
                     }
                     return filtered;
@@ -561,72 +476,104 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                     {errors.models}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-models-label">
-                    {" "}
-                    {inputPrompt.models}{" "}
-                  </FormHelperText>
+                  <FormHelperText id="helper-tex-channel-models-label"> {inputPrompt.models} </FormHelperText>
                 )}
               </FormControl>
               <Container
                 sx={{
-                  textAlign: "right",
+                  textAlign: 'right'
                 }}
               >
-                <ButtonGroup
-                  variant="outlined"
-                  aria-label="small outlined primary button group"
-                >
+                <ButtonGroup variant="outlined" aria-label="small outlined primary button group">
                   <Button
                     onClick={() => {
-                      setFieldValue("models", basicModels(values.type));
+                      setFieldValue('models', initialModel(basicModels));
                     }}
                   >
-                    填入渠道支持模型
+                    填入相关模型
                   </Button>
                   <Button
                     onClick={() => {
-                      setFieldValue("models", modelOptions);
+                      setFieldValue('models', modelOptions);
                     }}
                   >
                     填入所有模型
                   </Button>
                 </ButtonGroup>
               </Container>
-              <FormControl
-                fullWidth
-                error={Boolean(touched.key && errors.key)}
-                sx={{ ...theme.typography.otherInput }}
-              >
-                <InputLabel htmlFor="channel-key-label">
-                  {inputLabel.key}
-                </InputLabel>
-                <OutlinedInput
-                  id="channel-key-label"
-                  label={inputLabel.key}
-                  type="text"
-                  value={values.key}
-                  name="key"
-                  onBlur={handleBlur}
-                  onChange={handleChange}
-                  inputProps={{}}
-                  aria-describedby="helper-text-channel-key-label"
-                />
-                {touched.key && errors.key ? (
-                  <FormHelperText error id="helper-tex-channel-key-label">
-                    {errors.key}
-                  </FormHelperText>
-                ) : (
-                  <FormHelperText id="helper-tex-channel-key-label">
-                    {" "}
-                    {inputPrompt.key}{" "}
-                  </FormHelperText>
-                )}
-              </FormControl>
-              <FormControl
-                fullWidth
-                error={Boolean(touched.model_mapping && errors.model_mapping)}
-                sx={{ ...theme.typography.otherInput }}
-              >
+              {inputLabel.key && (
+                <>
+                  <FormControl fullWidth error={Boolean(touched.key && errors.key)} sx={{ ...theme.typography.otherInput }}>
+                    {!batchAdd ? (
+                      <>
+                        <InputLabel htmlFor="channel-key-label">{inputLabel.key}</InputLabel>
+                        <OutlinedInput
+                          id="channel-key-label"
+                          label={inputLabel.key}
+                          type="text"
+                          value={values.key}
+                          name="key"
+                          onBlur={handleBlur}
+                          onChange={handleChange}
+                          inputProps={{}}
+                          aria-describedby="helper-text-channel-key-label"
+                        />
+                      </>
+                    ) : (
+                      <TextField
+                        multiline
+                        id="channel-key-label"
+                        label={inputLabel.key}
+                        value={values.key}
+                        name="key"
+                        onBlur={handleBlur}
+                        onChange={handleChange}
+                        aria-describedby="helper-text-channel-key-label"
+                        minRows={5}
+                        placeholder={inputPrompt.key + '，一行一个密钥'}
+                      />
+                    )}
+
+                    {touched.key && errors.key ? (
+                      <FormHelperText error id="helper-tex-channel-key-label">
+                        {errors.key}
+                      </FormHelperText>
+                    ) : (
+                      <FormHelperText id="helper-tex-channel-key-label"> {inputPrompt.key} </FormHelperText>
+                    )}
+                  </FormControl>
+                  {channelId === 0 && (
+                    <Container
+                      sx={{
+                        textAlign: 'right'
+                      }}
+                    >
+                      <Switch checked={batchAdd} onChange={(e) => setBatchAdd(e.target.checked)} />
+                      批量添加
+                    </Container>
+                  )}
+                </>
+              )}
+
+              {inputLabel.config &&
+                Object.keys(inputLabel.config).map((configName) => {
+                  return (
+                    <FormControl key={'config.' + configName} fullWidth sx={{ ...theme.typography.otherInput }}>
+                      <TextField
+                        multiline
+                        key={'config.' + configName}
+                        name={'config.' + configName}
+                        value={values.config?.[configName] || ''}
+                        label={configName}
+                        placeholder={inputPrompt.config[configName]}
+                        onChange={handleChange}
+                      />
+                      <FormHelperText id={`helper-tex-config.${configName}-label`}> {inputPrompt.config[configName]} </FormHelperText>
+                    </FormControl>
+                  );
+                })}
+
+              <FormControl fullWidth error={Boolean(touched.model_mapping && errors.model_mapping)} sx={{ ...theme.typography.otherInput }}>
                 {/* <InputLabel htmlFor="channel-model_mapping-label">{inputLabel.model_mapping}</InputLabel> */}
                 <TextField
                   multiline
@@ -641,28 +588,38 @@ const EditModal = ({ open, channelId, onCancel, onOk }) => {
                   placeholder={inputPrompt.model_mapping}
                 />
                 {touched.model_mapping && errors.model_mapping ? (
-                  <FormHelperText
-                    error
-                    id="helper-tex-channel-model_mapping-label"
-                  >
+                  <FormHelperText error id="helper-tex-channel-model_mapping-label">
                     {errors.model_mapping}
                   </FormHelperText>
                 ) : (
-                  <FormHelperText id="helper-tex-channel-model_mapping-label">
-                    {" "}
-                    {inputPrompt.model_mapping}{" "}
+                  <FormHelperText id="helper-tex-channel-model_mapping-label"> {inputPrompt.model_mapping} </FormHelperText>
+                )}
+              </FormControl>
+              <FormControl fullWidth error={Boolean(touched.system_prompt && errors.system_prompt)} sx={{ ...theme.typography.otherInput }}>
+                {/* <InputLabel htmlFor="channel-model_mapping-label">{inputLabel.model_mapping}</InputLabel> */}
+                <TextField
+                  multiline
+                  id="channel-system_prompt-label"
+                  label={inputLabel.system_prompt}
+                  value={values.system_prompt}
+                  name="system_prompt"
+                  onBlur={handleBlur}
+                  onChange={handleChange}
+                  aria-describedby="helper-text-channel-system_prompt-label"
+                  minRows={5}
+                  placeholder={inputPrompt.system_prompt}
+                />
+                {touched.system_prompt && errors.system_prompt ? (
+                  <FormHelperText error id="helper-tex-channel-system_prompt-label">
+                    {errors.system_prompt}
                   </FormHelperText>
+                ) : (
+                  <FormHelperText id="helper-tex-channel-system_prompt-label"> {inputPrompt.system_prompt} </FormHelperText>
                 )}
               </FormControl>
               <DialogActions>
                 <Button onClick={onCancel}>取消</Button>
-                <Button
-                  disableElevation
-                  disabled={isSubmitting}
-                  type="submit"
-                  variant="contained"
-                  color="primary"
-                >
+                <Button disableElevation disabled={isSubmitting} type="submit" variant="contained" color="primary">
                   提交
                 </Button>
               </DialogActions>
@@ -680,5 +637,5 @@ EditModal.propTypes = {
   open: PropTypes.bool,
   channelId: PropTypes.number,
   onCancel: PropTypes.func,
-  onOk: PropTypes.func,
+  onOk: PropTypes.func
 };
diff --git a/web/berry/src/views/Channel/component/NameLabel.js b/web/berry/src/views/Channel/component/NameLabel.js
index 3dfd7824..742a66d9 100644
--- a/web/berry/src/views/Channel/component/NameLabel.js
+++ b/web/berry/src/views/Channel/component/NameLabel.js
@@ -1,20 +1,20 @@
-import PropTypes from "prop-types";
-import { Tooltip, Stack, Container } from "@mui/material";
-import Label from "ui-component/Label";
-import { styled } from "@mui/material/styles";
-import { showSuccess } from "utils/common";
+import PropTypes from 'prop-types';
+import { Tooltip, Stack, Container } from '@mui/material';
+import Label from 'ui-component/Label';
+import { styled } from '@mui/material/styles';
+import { showSuccess, copy } from 'utils/common';
 
 const TooltipContainer = styled(Container)({
-  maxHeight: "250px",
-  overflow: "auto",
-  "&::-webkit-scrollbar": {
-    width: "0px", // Set the width to 0 to hide the scrollbar
-  },
+  maxHeight: '250px',
+  overflow: 'auto',
+  '&::-webkit-scrollbar': {
+    width: '0px' // Set the width to 0 to hide the scrollbar
+  }
 });
 
 const NameLabel = ({ name, models }) => {
   let modelMap = [];
-  modelMap = models.split(",");
+  modelMap = models.split(',');
   modelMap.sort();
 
   return (
@@ -28,8 +28,7 @@ const NameLabel = ({ name, models }) => {
                   variant="ghost"
                   key={index}
                   onClick={() => {
-                    navigator.clipboard.writeText(item);
-                    showSuccess("复制模型名称成功！");
+                    copy(item, '模型名称');
                   }}
                 >
                   {item}
@@ -48,7 +47,7 @@ const NameLabel = ({ name, models }) => {
 
 NameLabel.propTypes = {
   name: PropTypes.string,
-  models: PropTypes.string,
+  models: PropTypes.string
 };
 
 export default NameLabel;
diff --git a/web/berry/src/views/Channel/component/TableRow.js b/web/berry/src/views/Channel/component/TableRow.js
index f7acb92e..525f9188 100644
--- a/web/berry/src/views/Channel/component/TableRow.js
+++ b/web/berry/src/views/Channel/component/TableRow.js
@@ -11,10 +11,7 @@ import {
   MenuItem,
   TableCell,
   IconButton,
-  FormControl,
-  InputLabel,
-  InputAdornment,
-  Input,
+  TextField,
   Dialog,
   DialogActions,
   DialogContent,
@@ -31,12 +28,7 @@ import ResponseTimeLabel from "./ResponseTimeLabel";
 import GroupLabel from "./GroupLabel";
 import NameLabel from "./NameLabel";
 
-import {
-  IconDotsVertical,
-  IconEdit,
-  IconTrash,
-  IconPencil,
-} from "@tabler/icons-react";
+import { IconDotsVertical, IconEdit, IconTrash } from "@tabler/icons-react";
 
 export default function ChannelTableRow({
   item,
@@ -79,11 +71,19 @@ export default function ChannelTableRow({
     }
   };
 
-  const handlePriority = async () => {
-    if (priorityValve === "" || priorityValve === item.priority) {
+  const handlePriority = async (event) => {
+    const currentValue = parseInt(event.target.value);
+    if (isNaN(currentValue) || currentValue === priorityValve) {
       return;
     }
-    await manageChannel(item.id, "priority", priorityValve);
+
+    if (currentValue < 0) {
+      showError("优先级不能小于 0");
+      return;
+    }
+
+    await manageChannel(item.id, "priority", currentValue);
+    setPriority(currentValue);
   };
 
   const handleResponseTime = async () => {
@@ -93,7 +93,7 @@ export default function ChannelTableRow({
         test_time: Date.now() / 1000,
         response_time: time * 1000,
       });
-      showInfo(`通道 ${item.name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
+      showInfo(`渠道 ${item.name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
     }
   };
 
@@ -170,9 +170,7 @@ export default function ChannelTableRow({
             handle_action={handleResponseTime}
           />
         </TableCell>
-        <TableCell>
-          {renderNumber(item.used_quota)}
-        </TableCell>
+        <TableCell>{renderNumber(item.used_quota)}</TableCell>
         <TableCell>
           <Tooltip
             title={"点击更新余额"}
@@ -183,27 +181,16 @@ export default function ChannelTableRow({
           </Tooltip>
         </TableCell>
         <TableCell>
-          <FormControl sx={{ m: 1, width: "70px" }} variant="standard">
-            <InputLabel htmlFor={`priority-${item.id}`}>优先级</InputLabel>
-            <Input
-              id={`priority-${item.id}`}
-              type="text"
-              value={priorityValve}
-              onChange={(e) => setPriority(e.target.value)}
-              sx={{ textAlign: "center" }}
-              endAdornment={
-                <InputAdornment position="end">
-                  <IconButton
-                    onClick={handlePriority}
-                    sx={{ color: "rgb(99, 115, 129)" }}
-                    size="small"
-                  >
-                    <IconPencil />
-                  </IconButton>
-                </InputAdornment>
-              }
-            />
-          </FormControl>
+          <TextField
+            id={`priority-${item.id}`}
+            onBlur={handlePriority}
+            type="number"
+            label="优先级"
+            variant="standard"
+            defaultValue={item.priority}
+            inputProps={{ min: "0" }}
+            sx={{ width: 80 }}
+          />
         </TableCell>
 
         <TableCell>
@@ -243,9 +230,9 @@ export default function ChannelTableRow({
       </Popover>
 
       <Dialog open={openDelete} onClose={handleDeleteClose}>
-        <DialogTitle>删除通道</DialogTitle>
+        <DialogTitle>删除渠道</DialogTitle>
         <DialogContent>
-          <DialogContentText>是否删除通道 {item.name}？</DialogContentText>
+          <DialogContentText>是否删除渠道 {item.name}？</DialogContentText>
         </DialogContent>
         <DialogActions>
           <Button onClick={handleDeleteClose}>关闭</Button>
@@ -281,6 +268,10 @@ function renderBalance(type, balance) {
       return <span>¥{balance.toFixed(2)}</span>;
     case 13: // AIGC2D
       return <span>{renderNumber(balance)}</span>;
+    case 36: // DeepSeek
+      return <span>¥{balance.toFixed(2)}</span>;
+    case 44: // SiliconFlow
+      return <span>¥{balance.toFixed(2)}</span>;
     default:
       return <span>不支持</span>;
   }
diff --git a/web/berry/src/views/Channel/index.js b/web/berry/src/views/Channel/index.js
index e801a8c3..3ca8fdef 100644
--- a/web/berry/src/views/Channel/index.js
+++ b/web/berry/src/views/Channel/index.js
@@ -1,5 +1,5 @@
 import { useState, useEffect } from 'react';
-import { showError, showSuccess, showInfo } from 'utils/common';
+import { showError, showSuccess, showInfo, loadChannelModels } from 'utils/common';
 
 import { useTheme } from '@mui/material/styles';
 import Table from '@mui/material/Table';
@@ -8,7 +8,6 @@ import TableContainer from '@mui/material/TableContainer';
 import PerfectScrollbar from 'react-perfect-scrollbar';
 import TablePagination from '@mui/material/TablePagination';
 import LinearProgress from '@mui/material/LinearProgress';
-import Alert from '@mui/material/Alert';
 import ButtonGroup from '@mui/material/ButtonGroup';
 import Toolbar from '@mui/material/Toolbar';
 import useMediaQuery from '@mui/material/useMediaQuery';
@@ -135,7 +134,7 @@ export default function ChannelPage() {
     const res = await API.get(`/api/channel/test`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已成功开始测试所有通道，请刷新页面查看结果。');
+      showInfo('已成功开始测试所有渠道，请刷新页面查看结果。');
     } else {
       showError(message);
     }
@@ -159,7 +158,7 @@ export default function ChannelPage() {
     const res = await API.get(`/api/channel/update_balance`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已更新完毕所有已启用通道余额！');
+      showInfo('已更新完毕所有已启用渠道余额！');
     } else {
       showError(message);
     }
@@ -189,6 +188,7 @@ export default function ChannelPage() {
       .catch((reason) => {
         showError(reason);
       });
+    loadChannelModels().then();
   }, []);
 
   return (
@@ -200,7 +200,7 @@ export default function ChannelPage() {
         </Button>
       </Stack>
       <Card>
-        <Box component="form" onSubmit={searchChannels} noValidate sx={{marginTop: 2}}>
+        <Box component="form" onSubmit={searchChannels} noValidate sx={{ marginTop: 2 }}>
           <TableToolBar filterName={searchKeyword} handleFilterName={handleSearchKeyword} placeholder={'搜索渠道的 ID，名称和密钥 ...'} />
         </Box>
         <Toolbar
@@ -214,7 +214,7 @@ export default function ChannelPage() {
         >
           <Container>
             {matchUpMd ? (
-              <ButtonGroup variant="outlined" aria-label="outlined small primary button group" sx={{marginBottom: 2}}>
+              <ButtonGroup variant="outlined" aria-label="outlined small primary button group" sx={{ marginBottom: 2 }}>
                 <Button onClick={handleRefresh} startIcon={<IconRefresh width={'18px'} />}>
                   刷新
                 </Button>
diff --git a/web/berry/src/views/Channel/type/Config.js b/web/berry/src/views/Channel/type/Config.js
index 7e42ca8d..67b90733 100644
--- a/web/berry/src/views/Channel/type/Config.js
+++ b/web/berry/src/views/Channel/type/Config.js
@@ -1,176 +1,232 @@
 const defaultConfig = {
   input: {
-    name: "",
+    name: '',
     type: 1,
-    key: "",
-    base_url: "",
-    other: "",
-    model_mapping: "",
+    key: '',
+    base_url: '',
+    other: '',
+    model_mapping: '',
     models: [],
-    groups: ["default"],
+    groups: ['default'],
+    config: {}
   },
   inputLabel: {
-    name: "渠道名称",
-    type: "渠道类型",
-    base_url: "渠道API地址",
-    key: "密钥",
-    other: "其他参数",
-    models: "模型",
-    model_mapping: "模型映射关系",
-    groups: "用户组",
+    name: '渠道名称',
+    type: '渠道类型',
+    base_url: '渠道API地址',
+    key: '密钥',
+    other: '其他参数',
+    models: '模型',
+    model_mapping: '模型映射关系',
+    system_prompt: '系统提示词',
+    groups: '用户组',
+    config: null
   },
   prompt: {
-    type: "请选择渠道类型",
-    name: "请为渠道命名",
-    base_url: "可空，请输入中转API地址，例如通过cloudflare中转",
-    key: "请输入渠道对应的鉴权密钥",
-    other: "",
-    models: "请选择该渠道所支持的模型",
+    type: '请选择渠道类型',
+    name: '请为渠道命名',
+    base_url: '可空，请输入中转API地址，例如通过cloudflare中转',
+    key: '请输入渠道对应的鉴权密钥',
+    other: '',
+    models: '请选择该渠道所支持的模型',
     model_mapping:
       '请输入要修改的模型映射关系，格式为：api请求模型ID:实际转发给渠道的模型ID，使用JSON数组表示，例如：{"gpt-3.5": "gpt-35"}',
-    groups: "请选择该渠道所支持的用户组",
+    system_prompt:"此项可选，用于强制设置给定的系统提示词，请配合自定义模型 & 模型重定向使用，首先创建一个唯一的自定义模型名称并在上面填入，之后将该自定义模型重定向映射到该渠道一个原生支持的模型此项可选，用于强制设置给定的系统提示词，请配合自定义模型 & 模型重定向使用，首先创建一个唯一的自定义模型名称并在上面填入，之后将该自定义模型重定向映射到该渠道一个原生支持的模型",
+    groups: '请选择该渠道所支持的用户组',
+    config: null
   },
-  modelGroup: "openai",
+  modelGroup: 'openai'
 };
 
 const typeConfig = {
   3: {
     inputLabel: {
-      base_url: "AZURE_OPENAI_ENDPOINT",
-      other: "默认 API 版本",
+      base_url: 'AZURE_OPENAI_ENDPOINT',
+      other: '默认 API 版本'
     },
     prompt: {
-      base_url: "请填写AZURE_OPENAI_ENDPOINT",
-      other: "请输入默认API版本，例如：2024-03-01-preview",
-    },
+      base_url: '请填写AZURE_OPENAI_ENDPOINT',
+      other: '请输入默认API版本，例如：2024-03-01-preview'
+    }
   },
   11: {
     input: {
-      models: ["PaLM-2"],
+      models: ['PaLM-2']
     },
-    modelGroup: "google palm",
+    modelGroup: 'google palm'
   },
   14: {
     input: {
-      models: ["claude-instant-1", "claude-2", "claude-2.0", "claude-2.1"],
+      models: ['claude-instant-1', 'claude-2', 'claude-2.0', 'claude-2.1']
     },
-    modelGroup: "anthropic",
+    modelGroup: 'anthropic'
   },
   15: {
     input: {
-      models: ["ERNIE-Bot", "ERNIE-Bot-turbo", "ERNIE-Bot-4", "Embedding-V1"],
+      models: ['ERNIE-Bot', 'ERNIE-Bot-turbo', 'ERNIE-Bot-4', 'Embedding-V1']
     },
     prompt: {
-      key: "按照如下格式输入：APIKey|SecretKey",
+      key: '按照如下格式输入：APIKey|SecretKey'
     },
-    modelGroup: "baidu",
+    modelGroup: 'baidu'
   },
   16: {
     input: {
-      models: ["glm-4", "glm-4v", "glm-3-turbo", "chatglm_turbo", "chatglm_pro", "chatglm_std", "chatglm_lite"],
+      models: ['glm-4', 'glm-4v', 'glm-3-turbo', 'chatglm_turbo', 'chatglm_pro', 'chatglm_std', 'chatglm_lite']
     },
-    modelGroup: "zhipu",
+    modelGroup: 'zhipu'
   },
   17: {
     inputLabel: {
-      other: "插件参数",
+      other: '插件参数'
     },
     input: {
-      models: [
-        "qwen-turbo",
-        "qwen-plus",
-        "qwen-max",
-        "qwen-max-longcontext",
-        "text-embedding-v1",
-      ],
+      models: ['qwen-turbo', 'qwen-plus', 'qwen-max', 'qwen-max-longcontext', 'text-embedding-v1']
     },
     prompt: {
-      other: "请输入插件参数，即 X-DashScope-Plugin 请求头的取值",
+      other: '请输入插件参数，即 X-DashScope-Plugin 请求头的取值'
     },
-    modelGroup: "ali",
+    modelGroup: 'ali'
   },
   18: {
     inputLabel: {
-      other: "版本号",
+      other: '版本号'
     },
     input: {
-      models: [
-          "SparkDesk",
-        'SparkDesk-v1.1',
-        'SparkDesk-v2.1',
-        'SparkDesk-v3.1',
-        'SparkDesk-v3.5'
-      ],
+      models: ['SparkDesk', 'SparkDesk-v1.1', 'SparkDesk-v2.1', 'SparkDesk-v3.1', 'SparkDesk-v3.1-128K', 'SparkDesk-v3.5', 'SparkDesk-v3.5-32K', 'SparkDesk-v4.0']
     },
     prompt: {
-      key: "按照如下格式输入：APPID|APISecret|APIKey",
-      other: "请输入版本号，例如：v3.1",
+      key: '按照如下格式输入：APPID|APISecret|APIKey',
+      other: '请输入版本号，例如：v3.1'
     },
-    modelGroup: "xunfei",
+    modelGroup: 'xunfei'
   },
   19: {
     input: {
-      models: [
-        "360GPT_S2_V9",
-        "embedding-bert-512-v1",
-        "embedding_s1_v1",
-        "semantic_similarity_s1_v1",
-      ],
+      models: ['360GPT_S2_V9', 'embedding-bert-512-v1', 'embedding_s1_v1', 'semantic_similarity_s1_v1']
     },
-    modelGroup: "360",
+    modelGroup: '360'
   },
   22: {
     prompt: {
-      key: "按照如下格式输入：APIKey-AppId，例如：fastgpt-0sp2gtvfdgyi4k30jwlgwf1i-64f335d84283f05518e9e041",
-    },
+      key: '按照如下格式输入：APIKey-AppId，例如：fastgpt-0sp2gtvfdgyi4k30jwlgwf1i-64f335d84283f05518e9e041'
+    }
   },
   23: {
     input: {
-      models: ["hunyuan"],
+      models: ['hunyuan']
     },
     prompt: {
-      key: "按照如下格式输入：AppId|SecretId|SecretKey",
+      key: '按照如下格式输入：AppId|SecretId|SecretKey'
     },
-    modelGroup: "tencent",
+    modelGroup: 'tencent'
   },
   24: {
     inputLabel: {
-      other: "版本号",
+      other: '版本号'
     },
     input: {
-      models: ["gemini-pro"],
+      models: ['gemini-pro']
     },
     prompt: {
-      other: "请输入版本号，例如：v1",
+      other: '请输入版本号，例如：v1'
     },
-    modelGroup: "google gemini",
+    modelGroup: 'google gemini'
   },
   25: {
     input: {
-      models: ['moonshot-v1-8k', 'moonshot-v1-32k', 'moonshot-v1-128k'],
+      models: ['moonshot-v1-8k', 'moonshot-v1-32k', 'moonshot-v1-128k']
     },
-    modelGroup: "moonshot",
+    modelGroup: 'moonshot'
   },
   26: {
     input: {
-      models: ['Baichuan2-Turbo', 'Baichuan2-Turbo-192k', 'Baichuan-Text-Embedding'],
+      models: ['Baichuan2-Turbo', 'Baichuan2-Turbo-192k', 'Baichuan-Text-Embedding']
     },
-    modelGroup: "baichuan",
+    modelGroup: 'baichuan'
   },
   27: {
     input: {
-      models: ['abab5.5s-chat', 'abab5.5-chat', 'abab6-chat'],
+      models: ['abab5.5s-chat', 'abab5.5-chat', 'abab6-chat']
     },
-    modelGroup: "minimax",
+    modelGroup: 'minimax'
   },
   29: {
-    modelGroup: "groq",
+    modelGroup: 'groq'
   },
   30: {
-    modelGroup: "ollama",
+    modelGroup: 'ollama'
   },
   31: {
-    modelGroup: "lingyiwanwu",
+    modelGroup: 'lingyiwanwu'
+  },
+  33: {
+    inputLabel: {
+      key: '',
+      config: {
+        region: 'Region',
+        ak: 'Access Key',
+        sk: 'Secret Key'
+      }
+    },
+    prompt: {
+      key: '',
+      config: {
+        region: 'region，e.g. us-west-2',
+        ak: 'AWS IAM Access Key',
+        sk: 'AWS IAM Secret Key'
+      }
+    },
+    modelGroup: 'anthropic'
+  },
+  37: {
+    inputLabel: {
+      config: {
+        user_id: 'Account ID'
+      }
+    },
+    prompt: {
+      config: {
+        user_id: '请输入 Account ID，例如：d8d7c61dbc334c32d3ced580e4bf42b4'
+      }
+    },
+    modelGroup: 'Cloudflare'
+  },
+  34: {
+    inputLabel: {
+      config: {
+        user_id: 'User ID'
+      }
+    },
+    prompt: {
+      models: '对于 Coze 而言，模型名称即 Bot ID，你可以添加一个前缀 `bot-`，例如：`bot-123456`',
+      config: {
+        user_id: '生成该密钥的用户 ID'
+      }
+    },
+    modelGroup: 'Coze'
+  },
+  42: {
+    inputLabel: {
+      key: '',
+      config: {
+        region: 'Vertex AI Region',
+        vertex_ai_project_id: 'Vertex AI Project ID',
+        vertex_ai_adc: 'Google Cloud Application Default Credentials JSON'
+      }
+    },
+    prompt: {
+      key: '',
+      config: {
+        region: 'Vertex AI Region.g. us-east5',
+        vertex_ai_project_id: 'Vertex AI Project ID',
+        vertex_ai_adc: 'Google Cloud Application Default Credentials JSON: https://cloud.google.com/docs/authentication/application-default-credentials'
+      }
+    },
+    modelGroup: 'anthropic'
+  },
+  45: {
+    modelGroup: 'xai'
   },
 };
 
diff --git a/web/berry/src/views/Dashboard/component/StatisticalLineChartCard.js b/web/berry/src/views/Dashboard/component/StatisticalLineChartCard.js
index 9daa9519..e6b46e25 100644
--- a/web/berry/src/views/Dashboard/component/StatisticalLineChartCard.js
+++ b/web/berry/src/views/Dashboard/component/StatisticalLineChartCard.js
@@ -12,7 +12,7 @@ import MainCard from 'ui-component/cards/MainCard';
 import SkeletonTotalOrderCard from 'ui-component/cards/Skeleton/EarningCard';
 
 const CardWrapper = styled(MainCard)(({ theme }) => ({
-  backgroundColor: theme.palette.primary.dark,
+  ...theme.typography.CardWrapper,
   color: '#fff',
   overflow: 'hidden',
   position: 'relative',
diff --git a/web/berry/src/views/Profile/index.js b/web/berry/src/views/Profile/index.js
index e0683228..b8fa0e7f 100644
--- a/web/berry/src/views/Profile/index.js
+++ b/web/berry/src/views/Profile/index.js
@@ -12,20 +12,23 @@ import {
   DialogTitle,
   DialogContent,
   DialogActions,
-  Divider
+  Divider,
+  SvgIcon
 } from '@mui/material';
 import Grid from '@mui/material/Unstable_Grid2';
 import SubCard from 'ui-component/cards/SubCard';
 import { IconBrandWechat, IconBrandGithub, IconMail } from '@tabler/icons-react';
 import Label from 'ui-component/Label';
 import { API } from 'utils/api';
-import { showError, showSuccess } from 'utils/common';
-import { onGitHubOAuthClicked } from 'utils/common';
+import { onOidcClicked, showError, showSuccess } from 'utils/common';
+import { onGitHubOAuthClicked, onLarkOAuthClicked, copy } from 'utils/common';
 import * as Yup from 'yup';
 import WechatModal from 'views/Authentication/AuthForms/WechatModal';
 import { useSelector } from 'react-redux';
 import EmailModal from './component/EmailModal';
 import Turnstile from 'react-turnstile';
+import { ReactComponent as Lark } from 'assets/images/icons/lark.svg';
+import { ReactComponent as OIDC } from 'assets/images/icons/oidc.svg';
 
 const validationSchema = Yup.object().shape({
   username: Yup.string().required('用户名 不能为空').min(3, '用户名 不能小于 3 个字符'),
@@ -88,8 +91,7 @@ export default function Profile() {
     const { success, message, data } = res.data;
     if (success) {
       setInputs((inputs) => ({ ...inputs, access_token: data }));
-      navigator.clipboard.writeText(data);
-      showSuccess(`令牌已重置并已复制到剪贴板`);
+      copy(data, '访问令牌');
     } else {
       showError(message);
     }
@@ -122,6 +124,15 @@ export default function Profile() {
     loadUser().then();
   }, [status]);
 
+  function getOidcId(){
+    if (!inputs.oidc_id) return '';
+    let oidc_id = inputs.oidc_id;
+    if (inputs.oidc_id.length > 8) {
+      oidc_id = inputs.oidc_id.slice(0, 6) + '...' + inputs.oidc_id.slice(-6);
+    }
+    return oidc_id;
+  }
+
   return (
     <>
       <UserCard>
@@ -137,6 +148,12 @@ export default function Profile() {
               <Label variant="ghost" color={inputs.email ? 'primary' : 'default'}>
                 <IconMail /> {inputs.email || '未绑定'}
               </Label>
+              <Label variant="ghost" color={inputs.lark_id ? 'primary' : 'default'}>
+                <SvgIcon component={Lark} inheritViewBox="0 0 24 24" /> {inputs.lark_id || '未绑定'}
+              </Label>
+              <Label variant="ghost" color={inputs.oidc_id ? 'primary' : 'default'}>
+                <SvgIcon component={OIDC} inheritViewBox="0 0 24 24" /> {getOidcId() || '未绑定'}
+              </Label>
             </Stack>
             <SubCard title="个人信息">
               <Grid container spacing={2}>
@@ -205,6 +222,20 @@ export default function Profile() {
                     </Button>
                   </Grid>
                 )}
+                {status.lark_client_id && !inputs.lark_id && (
+                  <Grid xs={12} md={4}>
+                    <Button variant="contained" onClick={() => onLarkOAuthClicked(status.lark_client_id)}>
+                      绑定 飞书 账号
+                    </Button>
+                  </Grid>
+                )}
+                {status.oidc && !inputs.oidc_id && (
+                  <Grid xs={12} md={4}>
+                    <Button variant="contained" onClick={() => onOidcClicked(status.oidc_authorization_endpoint,status.oidc_client_id,true)}>
+                      绑定 OIDC 账号
+                    </Button>
+                  </Grid>
+                )}
                 <Grid xs={12} md={4}>
                   <Button
                     variant="contained"
diff --git a/web/berry/src/views/Redemption/component/TableRow.js b/web/berry/src/views/Redemption/component/TableRow.js
index 68c9a505..380af037 100644
--- a/web/berry/src/views/Redemption/component/TableRow.js
+++ b/web/berry/src/views/Redemption/component/TableRow.js
@@ -18,7 +18,7 @@ import {
 
 import Label from 'ui-component/Label';
 import TableSwitch from 'ui-component/Switch';
-import { timestamp2string, renderQuota, showSuccess } from 'utils/common';
+import { timestamp2string, renderQuota, copy } from 'utils/common';
 
 import { IconDotsVertical, IconEdit, IconTrash } from '@tabler/icons-react';
 
@@ -83,8 +83,7 @@ export default function RedemptionTableRow({ item, manageRedemption, handleOpenM
               variant="contained"
               color="primary"
               onClick={() => {
-                navigator.clipboard.writeText(item.key);
-                showSuccess('已复制到剪贴板！');
+                copy(item.key, '兑换码');
               }}
             >
               复制
diff --git a/web/berry/src/views/Setting/component/OperationSetting.js b/web/berry/src/views/Setting/component/OperationSetting.js
index d91298b2..2bed715b 100644
--- a/web/berry/src/views/Setting/component/OperationSetting.js
+++ b/web/berry/src/views/Setting/component/OperationSetting.js
@@ -371,7 +371,7 @@ const OperationSetting = () => {
                 value={inputs.ChannelDisableThreshold}
                 onChange={handleInputChange}
                 label="最长响应时间"
-                placeholder="单位秒，当运行通道全部测试时，超过此时间将自动禁用通道"
+                placeholder="单位秒，当运行渠道全部测试时，超过此时间将自动禁用渠道"
                 disabled={loading}
               />
             </FormControl>
@@ -392,7 +392,7 @@ const OperationSetting = () => {
             </FormControl>
           </Stack>
           <FormControlLabel
-            label="失败时自动禁用通道"
+            label="失败时自动禁用渠道"
             control={
               <Checkbox
                 checked={inputs.AutomaticDisableChannelEnabled === "true"}
@@ -402,7 +402,7 @@ const OperationSetting = () => {
             }
           />
           <FormControlLabel
-            label="成功时自动启用通道"
+            label="成功时自动启用渠道"
             control={
               <Checkbox
                 checked={inputs.AutomaticEnableChannelEnabled === "true"}
diff --git a/web/berry/src/views/Setting/component/SystemSetting.js b/web/berry/src/views/Setting/component/SystemSetting.js
index fcf062a3..84e4f667 100644
--- a/web/berry/src/views/Setting/component/SystemSetting.js
+++ b/web/berry/src/views/Setting/component/SystemSetting.js
@@ -31,6 +31,15 @@ const SystemSetting = () => {
     GitHubOAuthEnabled: '',
     GitHubClientId: '',
     GitHubClientSecret: '',
+    LarkClientId: '',
+    LarkClientSecret: '',
+    OidcEnabled: '',
+    OidcWellKnown: '',
+    OidcClientId: '',
+    OidcClientSecret: '',
+    OidcAuthorizationEndpoint: '',
+    OidcTokenEndpoint: '',
+    OidcUserinfoEndpoint: '',
     Notice: '',
     SMTPServer: '',
     SMTPPort: '',
@@ -48,7 +57,9 @@ const SystemSetting = () => {
     TurnstileSecretKey: '',
     RegisterEnabled: '',
     EmailDomainRestrictionEnabled: '',
-    EmailDomainWhitelist: []
+    EmailDomainWhitelist: [],
+    MessagePusherAddress: '',
+    MessagePusherToken: ''
   });
   const [originInputs, setOriginInputs] = useState({});
   let [loading, setLoading] = useState(false);
@@ -90,6 +101,7 @@ const SystemSetting = () => {
       case 'TurnstileCheckEnabled':
       case 'EmailDomainRestrictionEnabled':
       case 'RegisterEnabled':
+      case 'OidcEnabled':
         value = inputs[key] === 'true' ? 'false' : 'true';
         break;
       default:
@@ -134,8 +146,19 @@ const SystemSetting = () => {
       name === 'WeChatAccountQRCodeImageURL' ||
       name === 'TurnstileSiteKey' ||
       name === 'TurnstileSecretKey' ||
-      name === 'EmailDomainWhitelist'
-    ) {
+      name === 'EmailDomainWhitelist' ||
+      name === 'MessagePusherAddress' ||
+      name === 'MessagePusherToken' ||
+      name === 'LarkClientId' ||
+      name === 'LarkClientSecret' ||
+      name === 'OidcClientId' ||
+      name === 'OidcClientSecret' ||
+      name === 'OidcWellKnown' ||
+      name === 'OidcAuthorizationEndpoint' ||
+      name === 'OidcTokenEndpoint' ||
+      name === 'OidcUserinfoEndpoint'
+    )
+    {
       setInputs((inputs) => ({ ...inputs, [name]: value }));
     } else {
       await updateOption(name, value);
@@ -199,6 +222,61 @@ const SystemSetting = () => {
     }
   };
 
+  const submitMessagePusher = async () => {
+    if (originInputs['MessagePusherAddress'] !== inputs.MessagePusherAddress) {
+      await updateOption('MessagePusherAddress', removeTrailingSlash(inputs.MessagePusherAddress));
+    }
+    if (originInputs['MessagePusherToken'] !== inputs.MessagePusherToken && inputs.MessagePusherToken !== '') {
+      await updateOption('MessagePusherToken', inputs.MessagePusherToken);
+    }
+  };
+
+  const submitLarkOAuth = async () => {
+    if (originInputs['LarkClientId'] !== inputs.LarkClientId) {
+      await updateOption('LarkClientId', inputs.LarkClientId);
+    }
+    if (originInputs['LarkClientSecret'] !== inputs.LarkClientSecret && inputs.LarkClientSecret !== '') {
+      await updateOption('LarkClientSecret', inputs.LarkClientSecret);
+    }
+  };
+
+  const submitOidc = async () => {
+    if (inputs.OidcWellKnown !== '') {
+      if (!inputs.OidcWellKnown.startsWith('http://') && !inputs.OidcWellKnown.startsWith('https://')) {
+        showError('Well-Known URL 必须以 http:// 或 https:// 开头');
+        return;
+      }
+      try {
+        const res = await API.get(inputs.OidcWellKnown);
+        inputs.OidcAuthorizationEndpoint = res.data['authorization_endpoint'];
+        inputs.OidcTokenEndpoint = res.data['token_endpoint'];
+        inputs.OidcUserinfoEndpoint = res.data['userinfo_endpoint'];
+        showSuccess('获取 OIDC 配置成功！');
+      } catch (err) {
+        showError("获取 OIDC 配置失败，请检查网络状况和 Well-Known URL 是否正确");
+      }
+    }
+
+    if (originInputs['OidcWellKnown'] !== inputs.OidcWellKnown) {
+      await updateOption('OidcWellKnown', inputs.OidcWellKnown);
+    }
+    if (originInputs['OidcClientId'] !== inputs.OidcClientId) {
+      await updateOption('OidcClientId', inputs.OidcClientId);
+    }
+    if (originInputs['OidcClientSecret'] !== inputs.OidcClientSecret && inputs.OidcClientSecret !== '') {
+      await updateOption('OidcClientSecret', inputs.OidcClientSecret);
+    }
+    if (originInputs['OidcAuthorizationEndpoint'] !== inputs.OidcAuthorizationEndpoint) {
+      await updateOption('OidcAuthorizationEndpoint', inputs.OidcAuthorizationEndpoint);
+    }
+    if (originInputs['OidcTokenEndpoint'] !== inputs.OidcTokenEndpoint) {
+      await updateOption('OidcTokenEndpoint', inputs.OidcTokenEndpoint);
+    }
+    if (originInputs['OidcUserinfoEndpoint'] !== inputs.OidcUserinfoEndpoint) {
+      await updateOption('OidcUserinfoEndpoint', inputs.OidcUserinfoEndpoint);
+    }
+  };
+
   return (
     <>
       <Stack spacing={2}>
@@ -265,6 +343,12 @@ const SystemSetting = () => {
                 control={<Checkbox checked={inputs.GitHubOAuthEnabled === 'true'} onChange={handleInputChange} name="GitHubOAuthEnabled" />}
               />
             </Grid>
+            <Grid xs={12} md={3}>
+              <FormControlLabel
+                label="允许通过 OIDC 登录 & 注册"
+                control={<Checkbox checked={inputs.OidcEnabled === 'true'} onChange={handleInputChange} name="OidcEnabled" />}
+              />
+            </Grid>
             <Grid xs={12} md={3}>
               <FormControlLabel
                 label="允许通过微信登录 & 注册"
@@ -473,6 +557,61 @@ const SystemSetting = () => {
             </Grid>
           </Grid>
         </SubCard>
+        <SubCard
+          title="配置飞书授权登录"
+          subTitle={
+            <span>
+              {' '}
+              用以支持通过飞书进行登录注册，
+              <a href="https://open.feishu.cn/app" target="_blank" rel="noreferrer">
+                点击此处
+              </a>
+              管理你的飞书应用
+            </span>
+          }
+        >
+          <Grid container spacing={{ xs: 3, sm: 2, md: 4 }}>
+            <Grid xs={12}>
+              <Alert severity="info" sx={{ wordWrap: 'break-word' }}>
+                主页链接填 <code>{inputs.ServerAddress}</code>
+                ，重定向 URL 填 <code>{`${inputs.ServerAddress}/oauth/lark`}</code>
+              </Alert>
+            </Grid>
+            <Grid xs={12} md={6}>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="LarkClientId">App ID</InputLabel>
+                <OutlinedInput
+                  id="LarkClientId"
+                  name="LarkClientId"
+                  value={inputs.LarkClientId || ''}
+                  onChange={handleInputChange}
+                  label="App ID"
+                  placeholder="输入 App ID"
+                  disabled={loading}
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={12} md={6}>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="LarkClientSecret">App Secret</InputLabel>
+                <OutlinedInput
+                  id="LarkClientSecret"
+                  name="LarkClientSecret"
+                  value={inputs.LarkClientSecret || ''}
+                  onChange={handleInputChange}
+                  label="App Secret"
+                  placeholder="敏感信息不会发送到前端显示"
+                  disabled={loading}
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={12}>
+              <Button variant="contained" onClick={submitLarkOAuth}>
+                保存飞书 OAuth 设置
+              </Button>
+            </Grid>
+          </Grid>
+        </SubCard>
         <SubCard
           title="配置 WeChat Server"
           subTitle={
@@ -535,6 +674,166 @@ const SystemSetting = () => {
             </Grid>
           </Grid>
         </SubCard>
+
+        <SubCard
+          title="配置 OIDC"
+          subTitle={
+            <span>
+              用以支持通过 OIDC 登录，例如 Okta、Auth0 等兼容 OIDC 协议的 IdP
+            </span>
+          }
+        >
+          <Grid container spacing={ { xs: 3, sm: 2, md: 4 } }>
+            <Grid xs={ 12 } md={ 12 }>
+              <Alert severity="info" sx={ { wordWrap: 'break-word' } }>
+                主页链接填 <code>{ inputs.ServerAddress }</code>
+                ，重定向 URL 填 <code>{ `${ inputs.ServerAddress }/oauth/oidc` }</code>
+              </Alert> <br />
+              <Alert severity="info" sx={ { wordWrap: 'break-word' } }>
+                若你的 OIDC Provider 支持 Discovery Endpoint，你可以仅填写 OIDC Well-Known URL，系统会自动获取 OIDC 配置
+              </Alert>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcClientId">Client ID</InputLabel>
+                <OutlinedInput
+                  id="OidcClientId"
+                  name="OidcClientId"
+                  value={ inputs.OidcClientId || '' }
+                  onChange={ handleInputChange }
+                  label="Client ID"
+                  placeholder="输入 OIDC 的 Client ID"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcClientSecret">Client Secret</InputLabel>
+                <OutlinedInput
+                  id="OidcClientSecret"
+                  name="OidcClientSecret"
+                  value={ inputs.OidcClientSecret || '' }
+                  onChange={ handleInputChange }
+                  label="Client Secret"
+                  placeholder="敏感信息不会发送到前端显示"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcWellKnown">Well-Known URL</InputLabel>
+                <OutlinedInput
+                  id="OidcWellKnown"
+                  name="OidcWellKnown"
+                  value={ inputs.OidcWellKnown || '' }
+                  onChange={ handleInputChange }
+                  label="Well-Known URL"
+                  placeholder="请输入 OIDC 的 Well-Known URL"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcAuthorizationEndpoint">Authorization Endpoint</InputLabel>
+                <OutlinedInput
+                  id="OidcAuthorizationEndpoint"
+                  name="OidcAuthorizationEndpoint"
+                  value={ inputs.OidcAuthorizationEndpoint || '' }
+                  onChange={ handleInputChange }
+                  label="Authorization Endpoint"
+                  placeholder="输入 OIDC 的 Authorization Endpoint"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcTokenEndpoint">Token Endpoint</InputLabel>
+                <OutlinedInput
+                  id="OidcTokenEndpoint"
+                  name="OidcTokenEndpoint"
+                  value={ inputs.OidcTokenEndpoint || '' }
+                  onChange={ handleInputChange }
+                  label="Token Endpoint"
+                  placeholder="输入 OIDC 的 Token Endpoint"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 } md={ 6 }>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="OidcUserinfoEndpoint">Userinfo Endpoint</InputLabel>
+                <OutlinedInput
+                  id="OidcUserinfoEndpoint"
+                  name="OidcUserinfoEndpoint"
+                  value={ inputs.OidcUserinfoEndpoint || '' }
+                  onChange={ handleInputChange }
+                  label="Userinfo Endpoint"
+                  placeholder="输入 OIDC 的 Userinfo Endpoint"
+                  disabled={ loading }
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={ 12 }>
+              <Button variant="contained" onClick={ submitOidc }>
+                保存 OIDC 设置
+              </Button>
+            </Grid>
+          </Grid>
+        </SubCard>
+
+        <SubCard
+          title="配置 Message Pusher"
+          subTitle={
+            <span>
+              用以推送报警信息，
+              <a href="https://github.com/songquanpeng/message-pusher" target="_blank" rel="noreferrer">
+                点击此处
+              </a>
+              了解 Message Pusher
+            </span>
+          }
+        >
+          <Grid container spacing={{ xs: 3, sm: 2, md: 4 }}>
+            <Grid xs={12} md={6}>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="MessagePusherAddress">Message Pusher 推送地址</InputLabel>
+                <OutlinedInput
+                  id="MessagePusherAddress"
+                  name="MessagePusherAddress"
+                  value={inputs.MessagePusherAddress || ''}
+                  onChange={handleInputChange}
+                  label="Message Pusher 推送地址"
+                  placeholder="例如：https://msgpusher.com/push/your_username"
+                  disabled={loading}
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={12} md={6}>
+              <FormControl fullWidth>
+                <InputLabel htmlFor="MessagePusherToken">Message Pusher 访问凭证</InputLabel>
+                <OutlinedInput
+                  id="MessagePusherToken"
+                  name="MessagePusherToken"
+                  type="password"
+                  value={inputs.MessagePusherToken || ''}
+                  onChange={handleInputChange}
+                  label="Message Pusher 访问凭证"
+                  placeholder="敏感信息不会发送到前端显示"
+                  disabled={loading}
+                />
+              </FormControl>
+            </Grid>
+            <Grid xs={12}>
+              <Button variant="contained" onClick={submitMessagePusher}>
+                保存 Message Pusher 设置
+              </Button>
+            </Grid>
+          </Grid>
+        </SubCard>
         <SubCard
           title="配置 Turnstile"
           subTitle={
diff --git a/web/berry/src/views/Token/component/EditModal.js b/web/berry/src/views/Token/component/EditModal.js
index cf65a96a..2685eab3 100644
--- a/web/berry/src/views/Token/component/EditModal.js
+++ b/web/berry/src/views/Token/component/EditModal.js
@@ -1,9 +1,9 @@
-import PropTypes from "prop-types";
-import * as Yup from "yup";
-import { Formik } from "formik";
-import { useTheme } from "@mui/material/styles";
-import { useState, useEffect } from "react";
-import dayjs from "dayjs";
+import PropTypes from 'prop-types';
+import * as Yup from 'yup';
+import { Formik } from 'formik';
+import { useTheme } from '@mui/material/styles';
+import { useState, useEffect } from 'react';
+import dayjs from 'dayjs';
 import {
   Dialog,
   DialogTitle,
@@ -16,53 +16,66 @@ import {
   InputLabel,
   OutlinedInput,
   InputAdornment,
+  Autocomplete,
+  Checkbox,
+  TextField,
   Switch,
-  FormHelperText,
-} from "@mui/material";
+  FormHelperText
+} from '@mui/material';
 
-import { AdapterDayjs } from "@mui/x-date-pickers/AdapterDayjs";
-import { LocalizationProvider } from "@mui/x-date-pickers/LocalizationProvider";
-import { DateTimePicker } from "@mui/x-date-pickers/DateTimePicker";
-import { renderQuotaWithPrompt, showSuccess, showError } from "utils/common";
-import { API } from "utils/api";
-require("dayjs/locale/zh-cn");
+import { AdapterDayjs } from '@mui/x-date-pickers/AdapterDayjs';
+import { LocalizationProvider } from '@mui/x-date-pickers/LocalizationProvider';
+import { DateTimePicker } from '@mui/x-date-pickers/DateTimePicker';
+import { renderQuotaWithPrompt, showSuccess, showError } from 'utils/common';
+import { API } from 'utils/api';
+import CheckBoxOutlineBlankIcon from '@mui/icons-material/CheckBoxOutlineBlank';
+import CheckBoxIcon from '@mui/icons-material/CheckBox';
+import { createFilterOptions } from '@mui/material/Autocomplete';
+require('dayjs/locale/zh-cn');
+const icon = <CheckBoxOutlineBlankIcon fontSize="small" />;
+const checkedIcon = <CheckBoxIcon fontSize="small" />;
+const filter = createFilterOptions();
 
 const validationSchema = Yup.object().shape({
   is_edit: Yup.boolean(),
-  name: Yup.string().required("名称 不能为空"),
-  remain_quota: Yup.number().min(0, "必须大于等于0"),
+  name: Yup.string().required('名称 不能为空'),
+  remain_quota: Yup.number().min(0, '必须大于等于0'),
   expired_time: Yup.number(),
-  unlimited_quota: Yup.boolean(),
+  unlimited_quota: Yup.boolean()
 });
 
 const originInputs = {
   is_edit: false,
-  name: "",
+  name: '',
   remain_quota: 0,
   expired_time: -1,
   unlimited_quota: false,
+  subnet: '',
+  models: []
 };
 
 const EditModal = ({ open, tokenId, onCancel, onOk }) => {
   const theme = useTheme();
   const [inputs, setInputs] = useState(originInputs);
+  const [modelOptions, setModelOptions] = useState([]);
 
   const submit = async (values, { setErrors, setStatus, setSubmitting }) => {
     setSubmitting(true);
 
     values.remain_quota = parseInt(values.remain_quota);
     let res;
+    let models = values.models.join(',');
     if (values.is_edit) {
-      res = await API.put(`/api/token/`, { ...values, id: parseInt(tokenId) });
+      res = await API.put(`/api/token/`, { ...values, id: parseInt(tokenId), models: models });
     } else {
-      res = await API.post(`/api/token/`, values);
+      res = await API.post(`/api/token/`, { ...values, models: models });
     }
     const { success, message } = res.data;
     if (success) {
       if (values.is_edit) {
-        showSuccess("令牌更新成功！");
+        showSuccess('令牌更新成功！');
       } else {
-        showSuccess("令牌创建成功，请在列表页面点击复制获取令牌！");
+        showSuccess('令牌创建成功，请在列表页面点击复制获取令牌！');
       }
       setSubmitting(false);
       setStatus({ success: true });
@@ -78,61 +91,55 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
     const { success, message, data } = res.data;
     if (success) {
       data.is_edit = true;
+      if (data.models === '') {
+        data.models = [];
+      } else {
+        data.models = data.models.split(',');
+      }
       setInputs(data);
     } else {
       showError(message);
     }
   };
+  const loadAvailableModels = async () => {
+    let res = await API.get(`/api/user/available_models`);
+    const { success, message, data } = res.data;
+    if (success) {
+      setModelOptions(data);
+    } else {
+      showError(message);
+    }
+  };
 
   useEffect(() => {
     if (tokenId) {
       loadToken().then();
     } else {
-      setInputs({...originInputs});
+      setInputs({ ...originInputs });
     }
+    loadAvailableModels().then();
   }, [tokenId]);
 
   return (
-    <Dialog open={open} onClose={onCancel} fullWidth maxWidth={"md"}>
+    <Dialog open={open} onClose={onCancel} fullWidth maxWidth={'md'}>
       <DialogTitle
         sx={{
-          margin: "0px",
+          margin: '0px',
           fontWeight: 700,
-          lineHeight: "1.55556",
-          padding: "24px",
-          fontSize: "1.125rem",
+          lineHeight: '1.55556',
+          padding: '24px',
+          fontSize: '1.125rem'
         }}
       >
-        {tokenId ? "编辑Token" : "新建Token"}
+        {tokenId ? '编辑令牌' : '新建令牌'}
       </DialogTitle>
       <Divider />
       <DialogContent>
-        <Alert severity="info">
-          注意，令牌的额度仅用于限制令牌本身的最大额度使用量，实际的使用受到账户的剩余额度限制。
-        </Alert>
-        <Formik
-          initialValues={inputs}
-          enableReinitialize
-          validationSchema={validationSchema}
-          onSubmit={submit}
-        >
-          {({
-            errors,
-            handleBlur,
-            handleChange,
-            handleSubmit,
-            touched,
-            values,
-            setFieldError,
-            setFieldValue,
-            isSubmitting,
-          }) => (
+        <Alert severity="info">注意，令牌的额度仅用于限制令牌本身的最大额度使用量，实际的使用受到账户的剩余额度限制。</Alert>
+        <Formik initialValues={inputs} enableReinitialize validationSchema={validationSchema} onSubmit={submit}>
+          {({ errors, handleBlur, handleChange, handleSubmit, touched, values, setFieldError, setFieldValue, isSubmitting }) => (
             <form noValidate onSubmit={handleSubmit}>
-              <FormControl
-                fullWidth
-                error={Boolean(touched.name && errors.name)}
-                sx={{ ...theme.typography.otherInput }}
-              >
+              <FormControl fullWidth error={Boolean(touched.name && errors.name)} sx={{ ...theme.typography.otherInput }}>
                 <InputLabel htmlFor="channel-name-label">名称</InputLabel>
                 <OutlinedInput
                   id="channel-name-label"
@@ -142,7 +149,7 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
                   name="name"
                   onBlur={handleBlur}
                   onChange={handleChange}
-                  inputProps={{ autoComplete: "name" }}
+                  inputProps={{ autoComplete: 'name' }}
                   aria-describedby="helper-text-channel-name-label"
                 />
                 {touched.name && errors.name && (
@@ -151,42 +158,99 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
                   </FormHelperText>
                 )}
               </FormControl>
+              <FormControl fullWidth sx={{ ...theme.typography.otherInput }}>
+                <Autocomplete
+                  multiple
+                  freeSolo
+                  id="channel-models-label"
+                  options={modelOptions}
+                  value={values.models}
+                  onChange={(e, value) => {
+                    const event = {
+                      target: {
+                        name: 'models',
+                        value: value
+                      }
+                    };
+                    handleChange(event);
+                  }}
+                  onBlur={handleBlur}
+                  // filterSelectedOptions
+                  disableCloseOnSelect
+                  renderInput={(params) => <TextField {...params} name="models" error={Boolean(errors.models)} label="模型范围" />}
+                  filterOptions={(options, params) => {
+                    const filtered = filter(options, params);
+                    const { inputValue } = params;
+                    const isExisting = options.some((option) => inputValue === option);
+                    if (inputValue !== '' && !isExisting) {
+                      filtered.push(inputValue);
+                    }
+                    return filtered;
+                  }}
+                  renderOption={(props, option, { selected }) => (
+                    <li {...props}>
+                      <Checkbox icon={icon} checkedIcon={checkedIcon} style={{ marginRight: 8 }} checked={selected} />
+                      {option}
+                    </li>
+                  )}
+                />
+                {errors.models ? (
+                  <FormHelperText error id="helper-tex-channel-models-label">
+                    {errors.models}
+                  </FormHelperText>
+                ) : (
+                  <FormHelperText id="helper-tex-channel-models-label">请选择允许使用的模型，留空则不进行限制</FormHelperText>
+                )}
+              </FormControl>
+              <FormControl fullWidth error={Boolean(touched.subnet && errors.subnet)} sx={{ ...theme.typography.otherInput }}>
+                <InputLabel htmlFor="channel-subnet-label">IP 限制</InputLabel>
+                <OutlinedInput
+                  id="channel-subnet-label"
+                  label="IP 限制"
+                  type="text"
+                  value={values.subnet}
+                  name="subnet"
+                  onBlur={handleBlur}
+                  onChange={handleChange}
+                  inputProps={{ autoComplete: 'subnet' }}
+                  aria-describedby="helper-text-channel-subnet-label"
+                />
+                {touched.subnet && errors.subnet ? (
+                  <FormHelperText error id="helper-tex-channel-subnet-label">
+                    {errors.subnet}
+                  </FormHelperText>
+                ) : (
+                  <FormHelperText id="helper-tex-channel-subnet-label">
+                    请输入允许访问的网段，例如：192.168.0.0/24，请使用英文逗号分隔多个网段
+                  </FormHelperText>
+                )}
+              </FormControl>
               {values.expired_time !== -1 && (
-                <FormControl
-                  fullWidth
-                  error={Boolean(touched.expired_time && errors.expired_time)}
-                  sx={{ ...theme.typography.otherInput }}
-                >
-                  <LocalizationProvider
-                    dateAdapter={AdapterDayjs}
-                    adapterLocale={"zh-cn"}
-                  >
+                <FormControl fullWidth error={Boolean(touched.expired_time && errors.expired_time)} sx={{ ...theme.typography.otherInput }}>
+                  <LocalizationProvider dateAdapter={AdapterDayjs} adapterLocale={'zh-cn'}>
                     <DateTimePicker
                       label="过期时间"
                       ampm={false}
                       value={dayjs.unix(values.expired_time)}
                       onError={(newError) => {
                         if (newError === null) {
-                          setFieldError("expired_time", null);
+                          setFieldError('expired_time', null);
                         } else {
-                          setFieldError("expired_time", "无效的日期");
+                          setFieldError('expired_time', '无效的日期');
                         }
                       }}
                       onChange={(newValue) => {
-                        setFieldValue("expired_time", newValue.unix());
+                        setFieldValue('expired_time', newValue.unix());
                       }}
                       slotProps={{
                         actionBar: {
-                          actions: ["today", "accept"],
-                        },
+                          actions: ['today', 'accept']
+                        }
                       }}
                     />
                   </LocalizationProvider>
                   {errors.expired_time && (
-                    <FormHelperText
-                      error
-                      id="helper-tex-channel-expired_time-label"
-                    >
+                    <FormHelperText error id="helper-tex-channel-expired_time-label">
                       {errors.expired_time}
                     </FormHelperText>
                   )}
@@ -196,35 +260,22 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
                 checked={values.expired_time === -1}
                 onClick={() => {
                   if (values.expired_time === -1) {
-                    setFieldValue(
-                      "expired_time",
-                      Math.floor(Date.now() / 1000)
-                    );
+                    setFieldValue('expired_time', Math.floor(Date.now() / 1000));
                   } else {
-                    setFieldValue("expired_time", -1);
+                    setFieldValue('expired_time', -1);
                   }
                 }}
-              />{" "}
+              />{' '}
               永不过期
-              <FormControl
-                fullWidth
-                error={Boolean(touched.remain_quota && errors.remain_quota)}
-                sx={{ ...theme.typography.otherInput }}
-              >
-                <InputLabel htmlFor="channel-remain_quota-label">
-                  额度
-                </InputLabel>
+              <FormControl fullWidth error={Boolean(touched.remain_quota && errors.remain_quota)} sx={{ ...theme.typography.otherInput }}>
+                <InputLabel htmlFor="channel-remain_quota-label">额度</InputLabel>
                 <OutlinedInput
                   id="channel-remain_quota-label"
                   label="额度"
                   type="number"
                   value={values.remain_quota}
                   name="remain_quota"
-                  endAdornment={
-                    <InputAdornment position="end">
-                      {renderQuotaWithPrompt(values.remain_quota)}
-                    </InputAdornment>
-                  }
+                  endAdornment={<InputAdornment position="end">{renderQuotaWithPrompt(values.remain_quota)}</InputAdornment>}
                   onBlur={handleBlur}
                   onChange={handleChange}
                   aria-describedby="helper-text-channel-remain_quota-label"
@@ -232,10 +283,7 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
                 />
 
                 {touched.remain_quota && errors.remain_quota && (
-                  <FormHelperText
-                    error
-                    id="helper-tex-channel-remain_quota-label"
-                  >
+                  <FormHelperText error id="helper-tex-channel-remain_quota-label">
                     {errors.remain_quota}
                   </FormHelperText>
                 )}
@@ -243,19 +291,13 @@ const EditModal = ({ open, tokenId, onCancel, onOk }) => {
               <Switch
                 checked={values.unlimited_quota === true}
                 onClick={() => {
-                  setFieldValue("unlimited_quota", !values.unlimited_quota);
+                  setFieldValue('unlimited_quota', !values.unlimited_quota);
                 }}
-              />{" "}
+              />{' '}
               无限额度
               <DialogActions>
                 <Button onClick={onCancel}>取消</Button>
-                <Button
-                  disableElevation
-                  disabled={isSubmitting}
-                  type="submit"
-                  variant="contained"
-                  color="primary"
-                >
+                <Button disableElevation disabled={isSubmitting} type="submit" variant="contained" color="primary">
                   提交
                 </Button>
               </DialogActions>
@@ -273,5 +315,5 @@ EditModal.propTypes = {
   open: PropTypes.bool,
   tokenId: PropTypes.number,
   onCancel: PropTypes.func,
-  onOk: PropTypes.func,
+  onOk: PropTypes.func
 };
diff --git a/web/berry/src/views/Token/component/TableRow.js b/web/berry/src/views/Token/component/TableRow.js
index 2753764c..4306be5e 100644
--- a/web/berry/src/views/Token/component/TableRow.js
+++ b/web/berry/src/views/Token/component/TableRow.js
@@ -20,7 +20,7 @@ import {
 } from '@mui/material';
 
 import TableSwitch from 'ui-component/Switch';
-import { renderQuota, showSuccess, timestamp2string } from 'utils/common';
+import { renderQuota, timestamp2string, copy } from 'utils/common';
 
 import { IconDotsVertical, IconEdit, IconTrash, IconCaretDownFilled } from '@tabler/icons-react';
 
@@ -28,11 +28,12 @@ const COPY_OPTIONS = [
   {
     key: 'next',
     text: 'ChatGPT Next',
-    url: 'https://chat.oneapi.pro/#/?settings={"key":"sk-{key}","url":"{serverAddress}"}',
+    url: 'https://app.nextchat.dev/#/?settings={"key":"sk-{key}","url":"{serverAddress}"}',
     encode: false
   },
   { key: 'ama', text: 'BotGem', url: 'ama://set-api-key?server={serverAddress}&key=sk-{key}', encode: true },
-  { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true }
+  { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true },
+  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"sk-{key}","baseURL":"{serverAddress}"}}}', encode: true }
 ];
 
 function replacePlaceholders(text, key, serverAddress) {
@@ -141,8 +142,7 @@ export default function TokensTableRow({ item, manageToken, handleOpenModal, set
     if (type === 'link') {
       window.open(text);
     } else {
-      navigator.clipboard.writeText(text);
-      showSuccess('已复制到剪贴板！');
+      copy(text);
     }
     handleCloseMenu();
   };
@@ -192,7 +192,7 @@ export default function TokensTableRow({ item, manageToken, handleOpenModal, set
               id={`switch-${item.id}`}
               checked={statusSwitch === 1}
               onChange={handleStatus}
-            // disabled={statusSwitch !== 1 && statusSwitch !== 2}
+              // disabled={statusSwitch !== 1 && statusSwitch !== 2}
             />
           </Tooltip>
         </TableCell>
@@ -211,8 +211,7 @@ export default function TokensTableRow({ item, manageToken, handleOpenModal, set
               <Button
                 color="primary"
                 onClick={() => {
-                  navigator.clipboard.writeText(`sk-${item.key}`);
-                  showSuccess('已复制到剪贴板！');
+                  copy(`sk-${item.key}`);
                 }}
               >
                 复制
@@ -222,7 +221,9 @@ export default function TokensTableRow({ item, manageToken, handleOpenModal, set
               </Button>
             </ButtonGroup>
             <ButtonGroup size="small" aria-label="split button">
-              <Button color="primary" onClick={(e) => handleCopy(COPY_OPTIONS[0], 'link')}>聊天</Button>
+              <Button color="primary" onClick={(e) => handleCopy(COPY_OPTIONS[0], 'link')}>
+                聊天
+              </Button>
               <Button size="small" onClick={(e) => handleOpenMenu(e, 'link')}>
                 <IconCaretDownFilled size={'16px'} />
               </Button>
diff --git a/web/berry/src/views/Topup/component/InviteCard.js b/web/berry/src/views/Topup/component/InviteCard.js
index a95f85e5..73c9670f 100644
--- a/web/berry/src/views/Topup/component/InviteCard.js
+++ b/web/berry/src/views/Topup/component/InviteCard.js
@@ -4,7 +4,7 @@ import SubCard from 'ui-component/cards/SubCard';
 import inviteImage from 'assets/images/invite/cwok_casual_19.webp';
 import { useState } from 'react';
 import { API } from 'utils/api';
-import { showError, showSuccess } from 'utils/common';
+import { showError, copy } from 'utils/common';
 
 const InviteCard = () => {
   const theme = useTheme();
@@ -12,8 +12,7 @@ const InviteCard = () => {
 
   const handleInviteUrl = async () => {
     if (inviteUl) {
-      navigator.clipboard.writeText(inviteUl);
-      showSuccess(`邀请链接已复制到剪切板`);
+      copy(inviteUl, '邀请链接');
       return;
     }
     const res = await API.get('/api/user/aff');
@@ -21,8 +20,7 @@ const InviteCard = () => {
     if (success) {
       let link = `${window.location.origin}/register?aff=${data}`;
       setInviteUrl(link);
-      navigator.clipboard.writeText(link);
-      showSuccess(`邀请链接已复制到剪切板`);
+      copy(link, '邀请链接');
     } else {
       showError(message);
     }
diff --git a/web/build.sh b/web/build.sh
old mode 100644
new mode 100755
diff --git a/web/default/package.json b/web/default/package.json
index 438f020c..ba45011f 100644
--- a/web/default/package.json
+++ b/web/default/package.json
@@ -18,7 +18,7 @@
   },
   "scripts": {
     "start": "react-scripts start",
-    "build": "react-scripts build && mv -f build ../build/default",
+    "build": "react-scripts build && rm -rf ../build/default && mv -f build ../build/default",
     "test": "react-scripts test",
     "eject": "react-scripts eject"
   },
diff --git a/web/default/public/index.html b/web/default/public/index.html
index b8e324d2..f22b8843 100644
--- a/web/default/public/index.html
+++ b/web/default/public/index.html
@@ -1,5 +1,5 @@
 <!DOCTYPE html>
-<html lang="en">
+<html lang="zh-CN">
   <head>
     <meta charset="utf-8" />
     <link rel="icon" href="logo.png" />
diff --git a/web/default/src/App.js b/web/default/src/App.js
index 13c884dc..4ece4eeb 100644
--- a/web/default/src/App.js
+++ b/web/default/src/App.js
@@ -24,6 +24,7 @@ import EditRedemption from './pages/Redemption/EditRedemption';
 import TopUp from './pages/TopUp';
 import Log from './pages/Log';
 import Chat from './pages/Chat';
+import LarkOAuth from './components/LarkOAuth';
 
 const Home = lazy(() => import('./pages/Home'));
 const About = lazy(() => import('./pages/About'));
@@ -239,6 +240,14 @@ function App() {
           </Suspense>
         }
       />
+      <Route
+        path='/oauth/lark'
+        element={
+          <Suspense fallback={<Loading></Loading>}>
+            <LarkOAuth />
+          </Suspense>
+        }
+      />
       <Route
         path='/setting'
         element={
diff --git a/web/default/src/components/ChannelsTable.js b/web/default/src/components/ChannelsTable.js
index 5fd6129a..e745814b 100644
--- a/web/default/src/components/ChannelsTable.js
+++ b/web/default/src/components/ChannelsTable.js
@@ -1,5 +1,5 @@
 import React, { useEffect, useState } from 'react';
-import { Button, Form, Input, Label, Message, Pagination, Popup, Table } from 'semantic-ui-react';
+import { Button, Dropdown, Form, Input, Label, Message, Pagination, Popup, Table } from 'semantic-ui-react';
 import { Link } from 'react-router-dom';
 import {
   API,
@@ -33,7 +33,7 @@ function renderType(type) {
     }
     type2label[0] = { value: 0, text: '未知类型', color: 'grey' };
   }
-  return <Label basic color={type2label[type]?.color}>{type2label[type]?.text}</Label>;
+  return <Label basic color={type2label[type]?.color}>{type2label[type] ? type2label[type].text : type}</Label>;
 }
 
 function renderBalance(type, balance) {
@@ -52,11 +52,21 @@ function renderBalance(type, balance) {
       return <span>¥{balance.toFixed(2)}</span>;
     case 13: // AIGC2D
       return <span>{renderNumber(balance)}</span>;
+    case 36: // DeepSeek
+      return <span>¥{balance.toFixed(2)}</span>;
+    case 44: // SiliconFlow
+      return <span>¥{balance.toFixed(2)}</span>;
     default:
       return <span>不支持</span>;
   }
 }
 
+function isShowDetail() {
+  return localStorage.getItem("show_detail") === "true";
+}
+
+const promptID = "detail"
+
 const ChannelsTable = () => {
   const [channels, setChannels] = useState([]);
   const [loading, setLoading] = useState(true);
@@ -64,19 +74,40 @@ const ChannelsTable = () => {
   const [searchKeyword, setSearchKeyword] = useState('');
   const [searching, setSearching] = useState(false);
   const [updatingBalance, setUpdatingBalance] = useState(false);
-  const [showPrompt, setShowPrompt] = useState(shouldShowPrompt("channel-test"));
+  const [showPrompt, setShowPrompt] = useState(shouldShowPrompt(promptID));
+  const [showDetail, setShowDetail] = useState(isShowDetail());
 
   const loadChannels = async (startIdx) => {
     const res = await API.get(`/api/channel/?p=${startIdx}`);
     const { success, message, data } = res.data;
     if (success) {
-      if (startIdx === 0) {
-        setChannels(data);
-      } else {
-        let newChannels = [...channels];
-        newChannels.splice(startIdx * ITEMS_PER_PAGE, data.length, ...data);
-        setChannels(newChannels);
-      }
+        let localChannels = data.map((channel) => {
+            if (channel.models === '') {
+                channel.models = [];
+                channel.test_model = "";
+            } else {
+                channel.models = channel.models.split(',');
+                if (channel.models.length > 0) {
+                    channel.test_model = channel.models[0];
+                }
+                channel.model_options = channel.models.map((model) => {
+                    return {
+                        key: model,
+                        text: model,
+                        value: model,
+                    }
+                })
+                console.log('channel', channel)
+            }
+            return channel;
+        });
+        if (startIdx === 0) {
+            setChannels(localChannels);
+        } else {
+            let newChannels = [...channels];
+            newChannels.splice(startIdx * ITEMS_PER_PAGE, data.length, ...localChannels);
+            setChannels(newChannels);
+        }
     } else {
       showError(message);
     }
@@ -98,6 +129,11 @@ const ChannelsTable = () => {
     await loadChannels(activePage - 1);
   };
 
+  const toggleShowDetail = () => {
+    setShowDetail(!showDetail);
+    localStorage.setItem("show_detail", (!showDetail).toString());
+  }
+
   useEffect(() => {
     loadChannels(0)
       .then()
@@ -225,26 +261,38 @@ const ChannelsTable = () => {
     setSearching(false);
   };
 
-  const testChannel = async (id, name, idx) => {
-    const res = await API.get(`/api/channel/test/${id}/`);
-    const { success, message, time } = res.data;
+  const switchTestModel = async (idx, model) => {
+    let newChannels = [...channels];
+    let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
+    newChannels[realIdx].test_model = model;
+    setChannels(newChannels);
+  };
+
+  const testChannel = async (id, name, idx, m) => {
+    const res = await API.get(`/api/channel/test/${id}?model=${m}`);
+    const { success, message, time, model } = res.data;
     if (success) {
       let newChannels = [...channels];
       let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
       newChannels[realIdx].response_time = time * 1000;
       newChannels[realIdx].test_time = Date.now() / 1000;
       setChannels(newChannels);
-      showInfo(`通道 ${name} 测试成功，耗时 ${time.toFixed(2)} 秒。`);
+      showInfo(`渠道 ${name} 测试成功，模型 ${model}，耗时 ${time.toFixed(2)} 秒。`);
     } else {
       showError(message);
     }
+    let newChannels = [...channels];
+    let realIdx = (activePage - 1) * ITEMS_PER_PAGE + idx;
+    newChannels[realIdx].response_time = time * 1000;
+    newChannels[realIdx].test_time = Date.now() / 1000;
+    setChannels(newChannels);
   };
 
   const testChannels = async (scope) => {
     const res = await API.get(`/api/channel/test?scope=${scope}`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已成功开始测试通道，请刷新页面查看结果。');
+      showInfo('已成功开始测试渠道，请刷新页面查看结果。');
     } else {
       showError(message);
     }
@@ -270,7 +318,7 @@ const ChannelsTable = () => {
       newChannels[realIdx].balance = balance;
       newChannels[realIdx].balance_updated_time = Date.now() / 1000;
       setChannels(newChannels);
-      showInfo(`通道 ${name} 余额更新成功！`);
+      showInfo(`渠道 ${name} 余额更新成功！`);
     } else {
       showError(message);
     }
@@ -281,7 +329,7 @@ const ChannelsTable = () => {
     const res = await API.get(`/api/channel/update_balance`);
     const { success, message } = res.data;
     if (success) {
-      showInfo('已更新完毕所有已启用通道余额！');
+      showInfo('已更新完毕所有已启用渠道余额！');
     } else {
       showError(message);
     }
@@ -330,11 +378,13 @@ const ChannelsTable = () => {
         showPrompt && (
           <Message onDismiss={() => {
             setShowPrompt(false);
-            setPromptShown("channel-test");
+            setPromptShown(promptID);
           }}>
             OpenAI 渠道已经不再支持通过 key 获取余额，因此余额显示为 0。对于支持的渠道类型，请点击余额进行刷新。
             <br/>
             渠道测试仅支持 chat 模型，优先使用 gpt-3.5-turbo，如果该模型不可用则使用你所配置的模型列表中的第一个模型。
+            <br/>
+            点击下方详情按钮可以显示余额以及设置额外的测试模型。
           </Message>
         )
       }
@@ -394,6 +444,7 @@ const ChannelsTable = () => {
               onClick={() => {
                 sortChannel('balance');
               }}
+              hidden={!showDetail}
             >
               余额
             </Table.HeaderCell>
@@ -405,6 +456,7 @@ const ChannelsTable = () => {
             >
               优先级
             </Table.HeaderCell>
+            <Table.HeaderCell hidden={!showDetail}>测试模型</Table.HeaderCell>
             <Table.HeaderCell>操作</Table.HeaderCell>
           </Table.Row>
         </Table.Header>
@@ -432,7 +484,7 @@ const ChannelsTable = () => {
                       basic
                     />
                   </Table.Cell>
-                  <Table.Cell>
+                  <Table.Cell hidden={!showDetail}>
                     <Popup
                       trigger={<span onClick={() => {
                         updateChannelBalance(channel.id, channel.name, idx);
@@ -459,13 +511,24 @@ const ChannelsTable = () => {
                       basic
                     />
                   </Table.Cell>
+                  <Table.Cell hidden={!showDetail}>
+                    <Dropdown
+                      placeholder='请选择测试模型'
+                      selection
+                      options={channel.model_options}
+                      defaultValue={channel.test_model}
+                      onChange={(event, data) => {
+                        switchTestModel(idx, data.value);
+                      }}
+                    />
+                  </Table.Cell>
                   <Table.Cell>
                     <div>
                       <Button
                         size={'small'}
                         positive
                         onClick={() => {
-                          testChannel(channel.id, channel.name, idx);
+                          testChannel(channel.id, channel.name, idx, channel.test_model);
                         }}
                       >
                         测试
@@ -527,7 +590,7 @@ const ChannelsTable = () => {
 
         <Table.Footer>
           <Table.Row>
-            <Table.HeaderCell colSpan='9'>
+            <Table.HeaderCell colSpan={showDetail ? "10" : "8"}>
               <Button size='small' as={Link} to='/channel/add' loading={loading}>
                 添加新的渠道
               </Button>
@@ -565,6 +628,7 @@ const ChannelsTable = () => {
                 }
               />
               <Button size='small' onClick={refresh} loading={loading}>刷新</Button>
+              <Button size='small' onClick={toggleShowDetail}>{showDetail ? "隐藏详情" : "详情"}</Button>
             </Table.HeaderCell>
           </Table.Row>
         </Table.Footer>
diff --git a/web/default/src/components/LarkOAuth.js b/web/default/src/components/LarkOAuth.js
new file mode 100644
index 00000000..bc2fb682
--- /dev/null
+++ b/web/default/src/components/LarkOAuth.js
@@ -0,0 +1,58 @@
+import React, { useContext, useEffect, useState } from 'react';
+import { Dimmer, Loader, Segment } from 'semantic-ui-react';
+import { useNavigate, useSearchParams } from 'react-router-dom';
+import { API, showError, showSuccess } from '../helpers';
+import { UserContext } from '../context/User';
+
+const LarkOAuth = () => {
+  const [searchParams, setSearchParams] = useSearchParams();
+
+  const [userState, userDispatch] = useContext(UserContext);
+  const [prompt, setPrompt] = useState('处理中...');
+  const [processing, setProcessing] = useState(true);
+
+  let navigate = useNavigate();
+
+  const sendCode = async (code, state, count) => {
+    const res = await API.get(`/api/oauth/lark?code=${code}&state=${state}`);
+    const { success, message, data } = res.data;
+    if (success) {
+      if (message === 'bind') {
+        showSuccess('绑定成功！');
+        navigate('/setting');
+      } else {
+        userDispatch({ type: 'login', payload: data });
+        localStorage.setItem('user', JSON.stringify(data));
+        showSuccess('登录成功！');
+        navigate('/');
+      }
+    } else {
+      showError(message);
+      if (count === 0) {
+        setPrompt(`操作失败，重定向至登录界面中...`);
+        navigate('/setting'); // in case this is failed to bind lark
+        return;
+      }
+      count++;
+      setPrompt(`出现错误，第 ${count} 次重试中...`);
+      await new Promise((resolve) => setTimeout(resolve, count * 2000));
+      await sendCode(code, state, count);
+    }
+  };
+
+  useEffect(() => {
+    let code = searchParams.get('code');
+    let state = searchParams.get('state');
+    sendCode(code, state, 0).then();
+  }, []);
+
+  return (
+    <Segment style={{ minHeight: '300px' }}>
+      <Dimmer active inverted>
+        <Loader size='large'>{prompt}</Loader>
+      </Dimmer>
+    </Segment>
+  );
+};
+
+export default LarkOAuth;
diff --git a/web/default/src/components/LoginForm.js b/web/default/src/components/LoginForm.js
index b48f64c4..71566ef8 100644
--- a/web/default/src/components/LoginForm.js
+++ b/web/default/src/components/LoginForm.js
@@ -3,7 +3,8 @@ import { Button, Divider, Form, Grid, Header, Image, Message, Modal, Segment } f
 import { Link, useNavigate, useSearchParams } from 'react-router-dom';
 import { UserContext } from '../context/User';
 import { API, getLogo, showError, showSuccess, showWarning } from '../helpers';
-import { onGitHubOAuthClicked } from './utils';
+import { onGitHubOAuthClicked, onLarkOAuthClicked } from './utils';
+import larkIcon from '../images/lark.svg';
 
 const LoginForm = () => {
   const [inputs, setInputs] = useState({
@@ -124,29 +125,52 @@ const LoginForm = () => {
             点击注册
           </Link>
         </Message>
-        {status.github_oauth || status.wechat_login ? (
+        {status.github_oauth || status.wechat_login || status.lark_client_id ? (
           <>
             <Divider horizontal>Or</Divider>
-            {status.github_oauth ? (
-              <Button
-                circular
-                color='black'
-                icon='github'
-                onClick={() => onGitHubOAuthClicked(status.github_client_id)}
-              />
-            ) : (
-              <></>
-            )}
-            {status.wechat_login ? (
-              <Button
-                circular
-                color='green'
-                icon='wechat'
-                onClick={onWeChatLoginClicked}
-              />
-            ) : (
-              <></>
-            )}
+            <div style={{ display: "flex", justifyContent: "center" }}>
+              {status.github_oauth ? (
+                <Button
+                  circular
+                  color='black'
+                  icon='github'
+                  onClick={() => onGitHubOAuthClicked(status.github_client_id)}
+                />
+              ) : (
+                <></>
+              )}
+              {status.wechat_login ? (
+                <Button
+                  circular
+                  color='green'
+                  icon='wechat'
+                  onClick={onWeChatLoginClicked}
+                />
+              ) : (
+                <></>
+              )}
+              {status.lark_client_id ? (
+                <div style={{
+                  background: "radial-gradient(circle, #FFFFFF, #FFFFFF, #00D6B9, #2F73FF, #0a3A9C)",
+                  width: "36px",
+                  height: "36px",
+                  borderRadius: "10em",
+                  display: "flex",
+                  cursor: "pointer"
+                }}
+                  onClick={() => onLarkOAuthClicked(status.lark_client_id)}
+                >
+                  <Image
+                    src={larkIcon}
+                    avatar
+                    style={{ width: "16px", height: "16px", cursor: "pointer", margin: "auto" }}
+                    onClick={() => onLarkOAuthClicked(status.lark_client_id)}
+                  />
+                </div>
+              ) : (
+                <></>
+              )}
+            </div>
           </>
         ) : (
           <></>
diff --git a/web/default/src/components/OperationSetting.js b/web/default/src/components/OperationSetting.js
index b823bb28..6356ac66 100644
--- a/web/default/src/components/OperationSetting.js
+++ b/web/default/src/components/OperationSetting.js
@@ -261,7 +261,7 @@ const OperationSetting = () => {
               value={inputs.ChannelDisableThreshold}
               type='number'
               min='0'
-              placeholder='单位秒，当运行通道全部测试时，超过此时间将自动禁用通道'
+              placeholder='单位秒，当运行渠道全部测试时，超过此时间将自动禁用渠道'
             />
             <Form.Input
               label='额度提醒阈值'
@@ -277,13 +277,13 @@ const OperationSetting = () => {
           <Form.Group inline>
             <Form.Checkbox
               checked={inputs.AutomaticDisableChannelEnabled === 'true'}
-              label='失败时自动禁用通道'
+              label='失败时自动禁用渠道'
               name='AutomaticDisableChannelEnabled'
               onChange={handleInputChange}
             />
             <Form.Checkbox
               checked={inputs.AutomaticEnableChannelEnabled === 'true'}
-              label='成功时自动启用通道'
+              label='成功时自动启用渠道'
               name='AutomaticEnableChannelEnabled'
               onChange={handleInputChange}
             />
diff --git a/web/default/src/components/PersonalSetting.js b/web/default/src/components/PersonalSetting.js
index 6baf1f35..0c968462 100644
--- a/web/default/src/components/PersonalSetting.js
+++ b/web/default/src/components/PersonalSetting.js
@@ -4,7 +4,7 @@ import { Link, useNavigate } from 'react-router-dom';
 import { API, copy, showError, showInfo, showNotice, showSuccess } from '../helpers';
 import Turnstile from 'react-turnstile';
 import { UserContext } from '../context/User';
-import { onGitHubOAuthClicked } from './utils';
+import { onGitHubOAuthClicked, onLarkOAuthClicked } from './utils';
 
 const PersonalSetting = () => {
   const [userState, userDispatch] = useContext(UserContext);
@@ -247,6 +247,11 @@ const PersonalSetting = () => {
           <Button onClick={()=>{onGitHubOAuthClicked(status.github_client_id)}}>绑定 GitHub 账号</Button>
         )
       }
+      {
+        status.lark_client_id && (
+          <Button onClick={()=>{onLarkOAuthClicked(status.lark_client_id)}}>绑定飞书账号</Button>
+        )
+      }
       <Button
         onClick={() => {
           setShowEmailBindModal(true);
diff --git a/web/default/src/components/SystemSetting.js b/web/default/src/components/SystemSetting.js
index 09b98665..7eeae92e 100644
--- a/web/default/src/components/SystemSetting.js
+++ b/web/default/src/components/SystemSetting.js
@@ -10,6 +10,8 @@ const SystemSetting = () => {
     GitHubOAuthEnabled: '',
     GitHubClientId: '',
     GitHubClientSecret: '',
+    LarkClientId: '',
+    LarkClientSecret: '',
     Notice: '',
     SMTPServer: '',
     SMTPPort: '',
@@ -109,6 +111,8 @@ const SystemSetting = () => {
       name === 'ServerAddress' ||
       name === 'GitHubClientId' ||
       name === 'GitHubClientSecret' ||
+      name === 'LarkClientId' ||
+      name === 'LarkClientSecret' ||
       name === 'WeChatServerAddress' ||
       name === 'WeChatServerToken' ||
       name === 'WeChatAccountQRCodeImageURL' ||
@@ -212,6 +216,18 @@ const SystemSetting = () => {
     }
   };
 
+   const submitLarkOAuth = async () => {
+    if (originInputs['LarkClientId'] !== inputs.LarkClientId) {
+      await updateOption('LarkClientId', inputs.LarkClientId);
+    }
+    if (
+      originInputs['LarkClientSecret'] !== inputs.LarkClientSecret &&
+      inputs.LarkClientSecret !== ''
+    ) {
+      await updateOption('LarkClientSecret', inputs.LarkClientSecret);
+    }
+  };
+
   const submitTurnstile = async () => {
     if (originInputs['TurnstileSiteKey'] !== inputs.TurnstileSiteKey) {
       await updateOption('TurnstileSiteKey', inputs.TurnstileSiteKey);
@@ -469,6 +485,44 @@ const SystemSetting = () => {
             保存 GitHub OAuth 设置
           </Form.Button>
           <Divider />
+          <Header as='h3'>
+            配置飞书授权登录
+            <Header.Subheader>
+              用以支持通过飞书进行登录注册，
+              <a href='https://open.feishu.cn/app' target='_blank'>
+                点击此处
+              </a>
+              管理你的飞书应用
+            </Header.Subheader>
+          </Header>
+          <Message>
+            主页链接填 <code>{inputs.ServerAddress}</code>
+            ，重定向 URL 填{' '}
+            <code>{`${inputs.ServerAddress}/oauth/lark`}</code>
+          </Message>
+          <Form.Group widths={3}>
+            <Form.Input
+              label='App ID'
+              name='LarkClientId'
+              onChange={handleInputChange}
+              autoComplete='new-password'
+              value={inputs.LarkClientId}
+              placeholder='输入 App ID'
+            />
+            <Form.Input
+              label='App Secret'
+              name='LarkClientSecret'
+              onChange={handleInputChange}
+              type='password'
+              autoComplete='new-password'
+              value={inputs.LarkClientSecret}
+              placeholder='敏感信息不会发送到前端显示'
+            />
+          </Form.Group>
+          <Form.Button onClick={submitLarkOAuth}>
+            保存飞书 OAuth 设置
+          </Form.Button>
+          <Divider />
           <Header as='h3'>
             配置 WeChat Server
             <Header.Subheader>
diff --git a/web/default/src/components/TokensTable.js b/web/default/src/components/TokensTable.js
index 19a688bb..64f77845 100644
--- a/web/default/src/components/TokensTable.js
+++ b/web/default/src/components/TokensTable.js
@@ -10,11 +10,14 @@ const COPY_OPTIONS = [
   { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'BotGem', value: 'ama' },
   { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 const OPEN_LINK_OPTIONS = [
+  { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'BotGem', value: 'ama' },
   { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 function renderTimestamp(timestamp) {
@@ -99,7 +102,7 @@ const TokensTable = () => {
     if (nextLink) {
       nextUrl = nextLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     } else {
-      nextUrl = `https://chat.oneapi.pro/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
+      nextUrl = `https://app.nextchat.dev/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     }
 
     let url;
@@ -113,6 +116,9 @@ const TokensTable = () => {
       case 'next':
         url = nextUrl;
         break;
+      case 'lobechat':
+        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
+        break;
       default:
         url = `sk-${key}`;
     }
@@ -141,7 +147,7 @@ const TokensTable = () => {
     if (chatLink) {
       defaultUrl = chatLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     } else {
-      defaultUrl = `https://chat.oneapi.pro/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
+      defaultUrl = `https://app.nextchat.dev/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
     }
     let url;
     switch (type) {
@@ -152,7 +158,11 @@ const TokensTable = () => {
       case 'opencat':
         url = `opencat://team/join?domain=${encodedServerAddress}&token=sk-${key}`;
         break;
-        
+
+      case 'lobechat':
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}/v1"}}}`;
+        break;
+
       default:
         url = defaultUrl;
     }
diff --git a/web/default/src/components/utils.js b/web/default/src/components/utils.js
index 5363ba5e..b711e274 100644
--- a/web/default/src/components/utils.js
+++ b/web/default/src/components/utils.js
@@ -17,4 +17,13 @@ export async function onGitHubOAuthClicked(github_client_id) {
   window.open(
     `https://github.com/login/oauth/authorize?client_id=${github_client_id}&state=${state}&scope=user:email`
   );
+}
+
+export async function onLarkOAuthClicked(lark_client_id) {
+  const state = await getOAuthState();
+  if (!state) return;
+  let redirect_uri = `${window.location.origin}/oauth/lark`;
+  window.open(
+    `https://open.feishu.cn/open-apis/authen/v1/index?redirect_uri=${redirect_uri}&app_id=${lark_client_id}&state=${state}`
+  );
 }
\ No newline at end of file
diff --git a/web/default/src/constants/channel.constants.js b/web/default/src/constants/channel.constants.js
index 8d536e58..61425508 100644
--- a/web/default/src/constants/channel.constants.js
+++ b/web/default/src/constants/channel.constants.js
@@ -1,33 +1,48 @@
 export const CHANNEL_OPTIONS = [
-  { key: 1, text: 'OpenAI', value: 1, color: 'green' },
-  { key: 14, text: 'Anthropic Claude', value: 14, color: 'black' },
-  { key: 3, text: 'Azure OpenAI', value: 3, color: 'olive' },
-  { key: 11, text: 'Google PaLM2', value: 11, color: 'orange' },
-  { key: 24, text: 'Google Gemini', value: 24, color: 'orange' },
-  { key: 28, text: 'Mistral AI', value: 28, color: 'orange' },
-  { key: 15, text: '百度文心千帆', value: 15, color: 'blue' },
-  { key: 17, text: '阿里通义千问', value: 17, color: 'orange' },
-  { key: 18, text: '讯飞星火认知', value: 18, color: 'blue' },
-  { key: 16, text: '智谱 ChatGLM', value: 16, color: 'violet' },
-  { key: 19, text: '360 智脑', value: 19, color: 'blue' },
-  { key: 25, text: 'Moonshot AI', value: 25, color: 'black' },
-  { key: 23, text: '腾讯混元', value: 23, color: 'teal' },
-  { key: 26, text: '百川大模型', value: 26, color: 'orange' },
-  { key: 27, text: 'MiniMax', value: 27, color: 'red' },
-  { key: 29, text: 'Groq', value: 29, color: 'orange' },
-  { key: 30, text: 'Ollama', value: 30, color: 'black' },
-  { key: 31, text: '零一万物', value: 31, color: 'green' },
-  { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
-  { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
-  { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },
-  { key: 20, text: '代理：OpenRouter', value: 20, color: 'black' },
-  { key: 2, text: '代理：API2D', value: 2, color: 'blue' },
-  { key: 5, text: '代理：OpenAI-SB', value: 5, color: 'brown' },
-  { key: 7, text: '代理：OhMyGPT', value: 7, color: 'purple' },
-  { key: 10, text: '代理：AI Proxy', value: 10, color: 'purple' },
-  { key: 4, text: '代理：CloseAI', value: 4, color: 'teal' },
-  { key: 6, text: '代理：OpenAI Max', value: 6, color: 'violet' },
-  { key: 9, text: '代理：AI.LS', value: 9, color: 'yellow' },
-  { key: 12, text: '代理：API2GPT', value: 12, color: 'blue' },
-  { key: 13, text: '代理：AIGC2D', value: 13, color: 'purple' }
-];
\ No newline at end of file
+    { key: 1, text: 'OpenAI', value: 1, color: 'green' },
+    { key: 14, text: 'Anthropic Claude', value: 14, color: 'black' },
+    { key: 33, text: 'AWS', value: 33, color: 'black' },
+    { key: 3, text: 'Azure OpenAI', value: 3, color: 'olive' },
+    { key: 11, text: 'Google PaLM2', value: 11, color: 'orange' },
+    { key: 24, text: 'Google Gemini', value: 24, color: 'orange' },
+    { key: 28, text: 'Mistral AI', value: 28, color: 'orange' },
+    { key: 41, text: 'Novita', value: 41, color: 'purple' },
+    { key: 40, text: '字节跳动豆包', value: 40, color: 'blue' },
+    { key: 15, text: '百度文心千帆', value: 15, color: 'blue' },
+    { key: 17, text: '阿里通义千问', value: 17, color: 'orange' },
+    { key: 18, text: '讯飞星火认知', value: 18, color: 'blue' },
+    { key: 16, text: '智谱 ChatGLM', value: 16, color: 'violet' },
+    { key: 19, text: '360 智脑', value: 19, color: 'blue' },
+    { key: 25, text: 'Moonshot AI', value: 25, color: 'black' },
+    { key: 23, text: '腾讯混元', value: 23, color: 'teal' },
+    { key: 26, text: '百川大模型', value: 26, color: 'orange' },
+    { key: 27, text: 'MiniMax', value: 27, color: 'red' },
+    { key: 29, text: 'Groq', value: 29, color: 'orange' },
+    { key: 30, text: 'Ollama', value: 30, color: 'black' },
+    { key: 31, text: '零一万物', value: 31, color: 'green' },
+    { key: 32, text: '阶跃星辰', value: 32, color: 'blue' },
+    { key: 34, text: 'Coze', value: 34, color: 'blue' },
+    { key: 35, text: 'Cohere', value: 35, color: 'blue' },
+    { key: 36, text: 'DeepSeek', value: 36, color: 'black' },
+    { key: 37, text: 'Cloudflare', value: 37, color: 'orange' },
+    { key: 38, text: 'DeepL', value: 38, color: 'black' },
+    { key: 39, text: 'together.ai', value: 39, color: 'blue' },
+    { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
+    { key: 43, text: 'Proxy', value: 43, color: 'blue' },
+    { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
+    { key: 45, text: 'xAI', value: 45, color: 'blue' },
+    { key: 46, text: 'Replicate', value: 46, color: 'blue' },
+    { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
+    { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
+    { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },
+    { key: 20, text: '代理：OpenRouter', value: 20, color: 'black' },
+    { key: 2, text: '代理：API2D', value: 2, color: 'blue' },
+    { key: 5, text: '代理：OpenAI-SB', value: 5, color: 'brown' },
+    { key: 7, text: '代理：OhMyGPT', value: 7, color: 'purple' },
+    { key: 10, text: '代理：AI Proxy', value: 10, color: 'purple' },
+    { key: 4, text: '代理：CloseAI', value: 4, color: 'teal' },
+    { key: 6, text: '代理：OpenAI Max', value: 6, color: 'violet' },
+    { key: 9, text: '代理：AI.LS', value: 9, color: 'yellow' },
+    { key: 12, text: '代理：API2GPT', value: 12, color: 'blue' },
+    { key: 13, text: '代理：AIGC2D', value: 13, color: 'purple' }
+];
diff --git a/web/default/src/images/lark.svg b/web/default/src/images/lark.svg
new file mode 100644
index 00000000..239e1bef
--- /dev/null
+++ b/web/default/src/images/lark.svg
@@ -0,0 +1 @@
+<svg width="48" height="48" viewBox="0 0 48 48" fill="none" xmlns="http://www.w3.org/2000/svg"><rect width="48" height="48" fill="white"/><path d="M25.9958 24.7255L26.084 24.6373C26.1481 24.5812 26.2042 24.5171 26.2603 24.461L26.3805 24.3407L26.7492 23.9801L27.2543 23.4831L27.6871 23.0583L28.0878 22.6575L28.5127 22.2407L28.8974 21.8559L29.4345 21.3189C29.5387 21.2227 29.6429 21.1185 29.7471 21.0223C29.9394 20.846 30.1479 20.6696 30.3482 20.5013C30.5406 20.349 30.733 20.2047 30.9334 20.0604C31.214 19.86 31.4945 19.6757 31.791 19.4993C32.0876 19.323 32.3842 19.1707 32.6808 19.0264C32.9693 18.8901 33.2579 18.7619 33.5545 18.6497C33.7228 18.5856 33.8912 18.5214 34.0514 18.4653L34.3079 18.3852C33.5625 15.4434 32.1918 12.6941 30.2921 10.3215C29.9234 9.86464 29.3703 9.6001 28.7852 9.6001H13.2671C13.1067 9.6001 12.9785 9.72832 12.9785 9.88864C12.9785 9.98482 13.0186 10.065 13.0987 10.1211C18.397 14.0247 22.7816 19.0104 25.9637 24.7575L25.9958 24.7255Z" fill="#00CEB1"/><path d="M19.8406 38.3843C27.5836 38.3843 34.7175 34.1921 38.4768 27.427C38.605 27.1946 38.7253 26.9621 38.8455 26.7296C38.6611 27.0743 38.4607 27.411 38.2363 27.7396C38.1562 27.8518 38.076 27.9641 37.9959 28.0682C37.8916 28.2045 37.7794 28.3328 37.6752 28.461C37.5871 28.5652 37.4989 28.6614 37.4027 28.7576C37.2184 28.95 37.018 29.1343 36.8176 29.3106C36.7054 29.4068 36.5931 29.503 36.4729 29.5912C36.3366 29.6954 36.1924 29.7996 36.0561 29.8958C35.9679 29.9599 35.8717 30.016 35.7835 30.0722C35.6954 30.1283 35.5912 30.1923 35.487 30.2405C35.2866 30.3527 35.0782 30.4569 34.8698 30.5531C34.6854 30.6332 34.5011 30.7134 34.3167 30.7775C34.1163 30.8496 33.9079 30.9218 33.6995 30.9779C33.3869 31.0661 33.0663 31.1302 32.7376 31.1863C32.5052 31.2264 32.2728 31.2504 32.0323 31.2665C31.7838 31.2825 31.5353 31.2905 31.2788 31.2825C30.9982 31.2745 30.7177 31.2584 30.4372 31.2264C30.2287 31.2023 30.0204 31.1703 29.812 31.1382C29.6276 31.1061 29.4513 31.0661 29.2669 31.026C29.1707 31.0019 29.0745 30.9779 28.9783 30.9539C28.7138 30.8817 28.4413 30.8096 28.1768 30.7294L27.784 30.6092C27.5836 30.5531 27.3913 30.4889 27.1909 30.4328C27.0306 30.3848 26.8702 30.3287 26.7099 30.2805C26.5576 30.2324 26.4054 30.1844 26.253 30.1283L25.9404 30.024L25.5637 29.8878L25.2992 29.7916C25.1229 29.7275 24.9465 29.6633 24.7702 29.5992L24.4656 29.487L24.0568 29.3267C23.9125 29.2706 23.7682 29.2145 23.6239 29.1584L23.3434 29.0462L22.9987 28.9019L22.7342 28.7897L22.4617 28.6694L22.2212 28.5652L22.0048 28.469L21.7884 28.3728L21.5639 28.2686L21.2754 28.1324L20.9788 27.9881L20.6582 27.8358L20.3856 27.6995C15.6324 25.3189 11.3361 22.1127 7.69699 18.2251C7.58473 18.1049 7.40041 18.1049 7.28815 18.2091C7.23205 18.2652 7.20001 18.3373 7.20001 18.4175L7.20805 32.1161V33.2303C7.20805 33.8795 7.52863 34.4807 8.06569 34.8414C11.5525 37.1579 15.6484 38.3923 19.8406 38.3843Z" fill="#3370FF"/><path d="M37.8772 27.4253L37.8691 27.4414L37.8772 27.4253L37.9493 27.2891C37.9252 27.3372 37.9012 27.3773 37.8772 27.4253Z" fill="#133C92"/><path d="M38.2137 26.8012L38.2297 26.7611L38.2377 26.7451L38.2137 26.8012Z" fill="#133C92"/><path d="M43.3814 19.0825C40.6802 17.76 37.5781 17.4714 34.6845 18.289C34.5642 18.321 34.444 18.3611 34.3238 18.3932L34.0673 18.4733C33.899 18.5295 33.7307 18.5935 33.5703 18.6577C33.2738 18.7699 32.9852 18.8982 32.6966 19.0344C32.392 19.1787 32.0955 19.339 31.8069 19.5073C31.5183 19.6756 31.2297 19.86 30.9492 20.0685C30.7488 20.2127 30.5565 20.357 30.3641 20.5093C30.1557 20.6776 29.9553 20.8459 29.7629 21.0303C29.6587 21.1265 29.5545 21.2227 29.4503 21.3269L28.9133 21.8639L28.5285 22.2487L28.1117 22.6655L27.7109 23.0662L27.2781 23.4911L26.7731 23.9881L26.4044 24.3487L26.2842 24.469C26.228 24.5251 26.1639 24.5892 26.1078 24.6453L26.0196 24.7335L25.8834 24.8617L25.7231 25.006C24.3925 26.2324 22.9016 27.2824 21.3065 28.1321L21.5951 28.2684L21.8195 28.3726L22.0439 28.4688L22.2603 28.5649L22.5008 28.6692L22.7733 28.7894L23.0378 28.9016L23.3825 29.0459L23.6631 29.1581C23.8073 29.2142 23.9516 29.2703 24.0959 29.3264L24.5047 29.4867L24.8093 29.599C24.9857 29.6631 25.162 29.7272 25.3383 29.7913L25.6028 29.8875L25.9796 30.0157L26.2922 30.12C26.4444 30.1761 26.5968 30.2242 26.7491 30.2722C26.9094 30.3283 27.0697 30.3765 27.23 30.4246C27.4224 30.4887 27.6228 30.5448 27.8231 30.6009L28.2159 30.7212C28.4804 30.8013 28.745 30.8734 29.0175 30.9456C29.1137 30.9696 29.2098 30.9937 29.306 31.0177C29.4904 31.0578 29.6667 31.0978 29.8511 31.1299C30.0595 31.17 30.2679 31.1941 30.4763 31.2181C30.7568 31.2502 31.0374 31.2662 31.3179 31.2742C31.5744 31.2742 31.8229 31.2742 32.0714 31.2582C32.3119 31.2421 32.5443 31.2181 32.7768 31.178C33.0974 31.1299 33.418 31.0578 33.7226 30.9776C33.931 30.9215 34.1394 30.8574 34.3398 30.7773C34.5242 30.7131 34.7085 30.633 34.8929 30.5528C35.1013 30.4566 35.3097 30.3524 35.5101 30.2402C35.6063 30.1841 35.7105 30.128 35.8067 30.0719C35.9028 30.0157 35.991 29.9596 36.0792 29.8955C36.2235 29.7993 36.3597 29.6952 36.496 29.5909C36.6162 29.5027 36.7284 29.4066 36.8407 29.3104C37.041 29.134 37.2414 28.9497 37.4258 28.7573C37.522 28.6611 37.6102 28.5649 37.6983 28.4607C37.8105 28.3325 37.9148 28.2042 38.019 28.068C38.0991 27.9638 38.1793 27.8515 38.2595 27.7393C38.4839 27.4187 38.6843 27.0901 38.8686 26.7374L39.077 26.3206L40.9527 22.5773V22.5853C41.5218 21.2868 42.3554 20.0925 43.3814 19.0825Z" fill="#133C92"/><path d="M37.9012 27.4346L37.8932 27.4506L37.9012 27.4346L37.9733 27.2983L37.9012 27.4346Z" fill="#133C92"/><path d="M38.2295 26.809L38.2455 26.7689L38.2536 26.7529C38.2536 26.7769 38.2375 26.793 38.2295 26.809Z" fill="#133C92"/></svg>
\ No newline at end of file
diff --git a/web/default/src/pages/Channel/EditChannel.js b/web/default/src/pages/Channel/EditChannel.js
index 330b8c8e..f10658c3 100644
--- a/web/default/src/pages/Channel/EditChannel.js
+++ b/web/default/src/pages/Channel/EditChannel.js
@@ -43,6 +43,7 @@ const EditChannel = () => {
     base_url: '',
     other: '',
     model_mapping: '',
+    system_prompt: '',
     models: [],
     groups: ['default']
   };
@@ -54,6 +55,14 @@ const EditChannel = () => {
   const [basicModels, setBasicModels] = useState([]);
   const [fullModels, setFullModels] = useState([]);
   const [customModel, setCustomModel] = useState('');
+  const [config, setConfig] = useState({
+    region: '',
+    sk: '',
+    ak: '',
+    user_id: '',
+    vertex_ai_project_id: '',
+    vertex_ai_adc: ''
+  });
   const handleInputChange = (e, { name, value }) => {
     setInputs((inputs) => ({ ...inputs, [name]: value }));
     if (name === 'type') {
@@ -65,6 +74,10 @@ const EditChannel = () => {
     }
   };
 
+  const handleConfigChange = (e, { name, value }) => {
+    setConfig((inputs) => ({ ...inputs, [name]: value }));
+  };
+
   const loadChannel = async () => {
     let res = await API.get(`/api/channel/${channelId}`);
     const { success, message, data } = res.data;
@@ -83,6 +96,10 @@ const EditChannel = () => {
         data.model_mapping = JSON.stringify(JSON.parse(data.model_mapping), null, 2);
       }
       setInputs(data);
+      if (data.config !== '') {
+        setConfig(JSON.parse(data.config));
+      }
+      setBasicModels(getChannelModels(data.type));
     } else {
       showError(message);
     }
@@ -99,9 +116,6 @@ const EditChannel = () => {
       }));
       setOriginModelOptions(localModelOptions);
       setFullModels(res.data.data.map((model) => model.id));
-      setBasicModels(res.data.data.filter((model) => {
-        return model.id.startsWith('gpt-3') || model.id.startsWith('text-');
-      }).map((model) => model.id));
     } catch (error) {
       showError(error.message);
     }
@@ -137,17 +151,27 @@ const EditChannel = () => {
   useEffect(() => {
     if (isEdit) {
       loadChannel().then();
+    } else {
+      let localModels = getChannelModels(inputs.type);
+      setBasicModels(localModels);
     }
     fetchModels().then();
     fetchGroups().then();
   }, []);
 
   const submit = async () => {
+    if (inputs.key === '') {
+      if (config.ak !== '' && config.sk !== '' && config.region !== '') {
+        inputs.key = `${config.ak}|${config.sk}|${config.region}`;
+      } else if (config.region !== '' && config.vertex_ai_project_id !== '' && config.vertex_ai_adc !== '') {
+        inputs.key = `${config.region}|${config.vertex_ai_project_id}|${config.vertex_ai_adc}`;
+      }
+    }
     if (!isEdit && (inputs.name === '' || inputs.key === '')) {
       showInfo('请填写渠道名称和渠道密钥！');
       return;
     }
-    if (inputs.models.length === 0) {
+    if (inputs.type !== 43 && inputs.models.length === 0) {
       showInfo('请至少选择一个模型！');
       return;
     }
@@ -155,19 +179,17 @@ const EditChannel = () => {
       showInfo('模型映射必须是合法的 JSON 格式！');
       return;
     }
-    let localInputs = inputs;
+    let localInputs = {...inputs};
     if (localInputs.base_url && localInputs.base_url.endsWith('/')) {
       localInputs.base_url = localInputs.base_url.slice(0, localInputs.base_url.length - 1);
     }
     if (localInputs.type === 3 && localInputs.other === '') {
       localInputs.other = '2024-03-01-preview';
     }
-    if (localInputs.type === 18 && localInputs.other === '') {
-      localInputs.other = 'v2.1';
-    }
     let res;
     localInputs.models = localInputs.models.join(',');
     localInputs.group = localInputs.groups.join(',');
+    localInputs.config = JSON.stringify(config);
     if (isEdit) {
       res = await API.put(`/api/channel/`, { ...localInputs, id: parseInt(channelId) });
     } else {
@@ -335,63 +357,181 @@ const EditChannel = () => {
               </Form.Field>
             )
           }
-          <Form.Field>
-            <Form.Dropdown
-              label='模型'
-              placeholder={'请选择该渠道所支持的模型'}
-              name='models'
-              required
-              fluid
-              multiple
-              search
-              onLabelClick={(e, { value }) => {copy(value).then()}}
-              selection
-              onChange={handleInputChange}
-              value={inputs.models}
-              autoComplete='new-password'
-              options={modelOptions}
-            />
-          </Form.Field>
-          <div style={{ lineHeight: '40px', marginBottom: '12px' }}>
-            <Button type={'button'} onClick={() => {
-              handleInputChange(null, { name: 'models', value: basicModels });
-            }}>填入基础模型</Button>
-            <Button type={'button'} onClick={() => {
-              handleInputChange(null, { name: 'models', value: fullModels });
-            }}>填入所有模型</Button>
-            <Button type={'button'} onClick={() => {
-              handleInputChange(null, { name: 'models', value: [] });
-            }}>清除所有模型</Button>
-            <Input
-              action={
-                <Button type={'button'} onClick={addCustomModel}>填入</Button>
-              }
-              placeholder='输入自定义模型名称'
-              value={customModel}
-              onChange={(e, { value }) => {
-                setCustomModel(value);
-              }}
-              onKeyDown={(e) => {
-                if (e.key === 'Enter') {
-                  addCustomModel();
-                  e.preventDefault();
-                }
-              }}
-            />
-          </div>
-          <Form.Field>
-            <Form.TextArea
-              label='模型重定向'
-              placeholder={`此项可选，用于修改请求体中的模型名称，为一个 JSON 字符串，键为请求中模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
-              name='model_mapping'
-              onChange={handleInputChange}
-              value={inputs.model_mapping}
-              style={{ minHeight: 150, fontFamily: 'JetBrains Mono, Consolas' }}
-              autoComplete='new-password'
-            />
-          </Form.Field>
           {
-            batch ? <Form.Field>
+            inputs.type === 34 && (
+              <Message>
+                对于 Coze 而言，模型名称即 Bot ID，你可以添加一个前缀 `bot-`，例如：`bot-123456`。
+              </Message>
+            )
+          }
+          {
+            inputs.type === 40 && (
+              <Message>
+                对于豆包而言，需要手动去 <a target="_blank" href="https://console.volcengine.com/ark/region:ark+cn-beijing/endpoint">模型推理页面</a> 创建推理接入点，以接入点名称作为模型名称，例如：`ep-20240608051426-tkxvl`。
+              </Message>
+            )
+          }
+          {
+            inputs.type !== 43 && (
+              <Form.Field>
+                <Form.Dropdown
+                  label='模型'
+                  placeholder={'请选择该渠道所支持的模型'}
+                  name='models'
+                  required
+                  fluid
+                  multiple
+                  search
+                  onLabelClick={(e, { value }) => {
+                    copy(value).then();
+                  }}
+                  selection
+                  onChange={handleInputChange}
+                  value={inputs.models}
+                  autoComplete='new-password'
+                  options={modelOptions}
+                />
+              </Form.Field>
+            )
+          }
+          {
+            inputs.type !== 43 && (
+              <div style={{ lineHeight: '40px', marginBottom: '12px' }}>
+                <Button type={'button'} onClick={() => {
+                  handleInputChange(null, { name: 'models', value: basicModels });
+                }}>填入相关模型</Button>
+                <Button type={'button'} onClick={() => {
+                  handleInputChange(null, { name: 'models', value: fullModels });
+                }}>填入所有模型</Button>
+                <Button type={'button'} onClick={() => {
+                  handleInputChange(null, { name: 'models', value: [] });
+                }}>清除所有模型</Button>
+                <Input
+                  action={
+                    <Button type={'button'} onClick={addCustomModel}>填入</Button>
+                  }
+                  placeholder='输入自定义模型名称'
+                  value={customModel}
+                  onChange={(e, { value }) => {
+                    setCustomModel(value);
+                  }}
+                  onKeyDown={(e) => {
+                    if (e.key === 'Enter') {
+                      addCustomModel();
+                      e.preventDefault();
+                    }
+                  }}
+                />
+              </div>
+            )
+          }
+          {
+          inputs.type !== 43 && (<>
+              <Form.Field>
+                <Form.TextArea
+                  label='模型重定向'
+                  placeholder={`此项可选，用于修改请求体中的模型名称，为一个 JSON 字符串，键为请求中模型名称，值为要替换的模型名称，例如：\n${JSON.stringify(MODEL_MAPPING_EXAMPLE, null, 2)}`}
+                  name='model_mapping'
+                  onChange={handleInputChange}
+                  value={inputs.model_mapping}
+                  style={{ minHeight: 150, fontFamily: 'JetBrains Mono, Consolas' }}
+                  autoComplete='new-password'
+                />
+              </Form.Field>
+            <Form.Field>
+                <Form.TextArea
+                  label='系统提示词'
+                  placeholder={`此项可选，用于强制设置给定的系统提示词，请配合自定义模型 & 模型重定向使用，首先创建一个唯一的自定义模型名称并在上面填入，之后将该自定义模型重定向映射到该渠道一个原生支持的模型`}
+                  name='system_prompt'
+                  onChange={handleInputChange}
+                  value={inputs.system_prompt}
+                  style={{ minHeight: 150, fontFamily: 'JetBrains Mono, Consolas' }}
+                  autoComplete='new-password'
+                />
+              </Form.Field>
+              </>
+            )
+          }
+          {
+            inputs.type === 33 && (
+              <Form.Field>
+                <Form.Input
+                  label='Region'
+                  name='region'
+                  required
+                  placeholder={'region，e.g. us-west-2'}
+                  onChange={handleConfigChange}
+                  value={config.region}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='AK'
+                  name='ak'
+                  required
+                  placeholder={'AWS IAM Access Key'}
+                  onChange={handleConfigChange}
+                  value={config.ak}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='SK'
+                  name='sk'
+                  required
+                  placeholder={'AWS IAM Secret Key'}
+                  onChange={handleConfigChange}
+                  value={config.sk}
+                  autoComplete=''
+                />
+              </Form.Field>
+            )
+          }
+          {
+            inputs.type === 42 && (
+              <Form.Field>
+                <Form.Input
+                  label='Region'
+                  name='region'
+                  required
+                  placeholder={'Vertex AI Region.g. us-east5'}
+                  onChange={handleConfigChange}
+                  value={config.region}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='Vertex AI Project ID'
+                  name='vertex_ai_project_id'
+                  required
+                  placeholder={'Vertex AI Project ID'}
+                  onChange={handleConfigChange}
+                  value={config.vertex_ai_project_id}
+                  autoComplete=''
+                />
+                <Form.Input
+                  label='Google Cloud Application Default Credentials JSON'
+                  name='vertex_ai_adc'
+                  required
+                  placeholder={'Google Cloud Application Default Credentials JSON'}
+                  onChange={handleConfigChange}
+                  value={config.vertex_ai_adc}
+                  autoComplete=''
+                />
+              </Form.Field>
+            )
+          }
+          {
+            inputs.type === 34 && (
+              <Form.Input
+                label='User ID'
+                name='user_id'
+                required
+                placeholder={'生成该密钥的用户 ID'}
+                onChange={handleConfigChange}
+                value={config.user_id}
+                autoComplete=''
+              />)
+          }
+          {
+            inputs.type !== 33 && inputs.type !== 42 && (batch ? <Form.Field>
               <Form.TextArea
                 label='密钥'
                 name='key'
@@ -412,10 +552,25 @@ const EditChannel = () => {
                 value={inputs.key}
                 autoComplete='new-password'
               />
-            </Form.Field>
+            </Form.Field>)
           }
           {
-            !isEdit && (
+            inputs.type === 37 && (
+              <Form.Field>
+                <Form.Input
+                  label='Account ID'
+                  name='user_id'
+                  required
+                  placeholder={'请输入 Account ID，例如：d8d7c61dbc334c32d3ced580e4bf42b4'}
+                  onChange={handleConfigChange}
+                  value={config.user_id}
+                  autoComplete=''
+                />
+              </Form.Field>
+            )
+          }
+          {
+            inputs.type !== 33 && !isEdit && (
               <Form.Checkbox
                 checked={batch}
                 label='批量创建'
@@ -425,7 +580,7 @@ const EditChannel = () => {
             )
           }
           {
-            inputs.type !== 3 && inputs.type !== 8 && inputs.type !== 22 && (
+            inputs.type !== 3 && inputs.type !== 33 && inputs.type !== 8 && inputs.type !== 22 && (
               <Form.Field>
                 <Form.Input
                   label='代理'
diff --git a/web/default/src/pages/Channel/index.js b/web/default/src/pages/Channel/index.js
index 57b64893..edf668b3 100644
--- a/web/default/src/pages/Channel/index.js
+++ b/web/default/src/pages/Channel/index.js
@@ -2,7 +2,7 @@ import React from 'react';
 import { Header, Segment } from 'semantic-ui-react';
 import ChannelsTable from '../../components/ChannelsTable';
 
-const File = () => (
+const Channel = () => (
   <>
     <Segment>
       <Header as='h3'>管理渠道</Header>
@@ -11,4 +11,4 @@ const File = () => (
   </>
 );
 
-export default File;
+export default Channel;
diff --git a/web/default/src/pages/Token/EditToken.js b/web/default/src/pages/Token/EditToken.js
index 0ab37c29..684b7e92 100644
--- a/web/default/src/pages/Token/EditToken.js
+++ b/web/default/src/pages/Token/EditToken.js
@@ -1,19 +1,22 @@
 import React, { useEffect, useState } from 'react';
 import { Button, Form, Header, Message, Segment } from 'semantic-ui-react';
-import { useParams, useNavigate } from 'react-router-dom';
-import { API, showError, showSuccess, timestamp2string } from '../../helpers';
-import { renderQuota, renderQuotaWithPrompt } from '../../helpers/render';
+import { useNavigate, useParams } from 'react-router-dom';
+import { API, copy, showError, showSuccess, timestamp2string } from '../../helpers';
+import { renderQuotaWithPrompt } from '../../helpers/render';
 
 const EditToken = () => {
   const params = useParams();
   const tokenId = params.id;
   const isEdit = tokenId !== undefined;
   const [loading, setLoading] = useState(isEdit);
+  const [modelOptions, setModelOptions] = useState([]);
   const originInputs = {
     name: '',
     remain_quota: isEdit ? 0 : 500000,
     expired_time: -1,
-    unlimited_quota: false
+    unlimited_quota: false,
+    models: [],
+    subnet: "",
   };
   const [inputs, setInputs] = useState(originInputs);
   const { name, remain_quota, expired_time, unlimited_quota } = inputs;
@@ -22,8 +25,8 @@ const EditToken = () => {
     setInputs((inputs) => ({ ...inputs, [name]: value }));
   };
   const handleCancel = () => {
-    navigate("/token");
-  }
+    navigate('/token');
+  };
   const setExpiredTime = (month, day, hour, minute) => {
     let now = new Date();
     let timestamp = now.getTime() / 1000;
@@ -50,6 +53,11 @@ const EditToken = () => {
       if (data.expired_time !== -1) {
         data.expired_time = timestamp2string(data.expired_time);
       }
+      if (data.models === '') {
+        data.models = [];
+      } else {
+        data.models = data.models.split(',');
+      }
       setInputs(data);
     } else {
       showError(message);
@@ -60,8 +68,26 @@ const EditToken = () => {
     if (isEdit) {
       loadToken().then();
     }
+    loadAvailableModels().then();
   }, []);
 
+  const loadAvailableModels = async () => {
+    let res = await API.get(`/api/user/available_models`);
+    const { success, message, data } = res.data;
+    if (success) {
+      let options = data.map((model) => {
+        return {
+          key: model,
+          text: model,
+          value: model
+        };
+      });
+      setModelOptions(options);
+    } else {
+      showError(message);
+    }
+  };
+
   const submit = async () => {
     if (!isEdit && inputs.name === '') return;
     let localInputs = inputs;
@@ -74,6 +100,7 @@ const EditToken = () => {
       }
       localInputs.expired_time = Math.ceil(time / 1000);
     }
+    localInputs.models = localInputs.models.join(',');
     let res;
     if (isEdit) {
       res = await API.put(`/api/token/`, { ...localInputs, id: parseInt(tokenId) });
@@ -109,6 +136,34 @@ const EditToken = () => {
               required={!isEdit}
             />
           </Form.Field>
+          <Form.Field>
+            <Form.Dropdown
+              label='模型范围'
+              placeholder={'请选择允许使用的模型，留空则不进行限制'}
+              name='models'
+              fluid
+              multiple
+              search
+              onLabelClick={(e, { value }) => {
+                copy(value).then();
+              }}
+              selection
+              onChange={handleInputChange}
+              value={inputs.models}
+              autoComplete='new-password'
+              options={modelOptions}
+            />
+          </Form.Field>
+          <Form.Field>
+            <Form.Input
+              label='IP 限制'
+              name='subnet'
+              placeholder={'请输入允许访问的网段，例如：192.168.0.0/24，请使用英文逗号分隔多个网段'}
+              onChange={handleInputChange}
+              value={inputs.subnet}
+              autoComplete='new-password'
+            />
+          </Form.Field>
           <Form.Field>
             <Form.Input
               label='过期时间'
diff --git a/web/default/src/pages/TopUp/index.js b/web/default/src/pages/TopUp/index.js
index f52cb8d5..2fcf0eae 100644
--- a/web/default/src/pages/TopUp/index.js
+++ b/web/default/src/pages/TopUp/index.js
@@ -8,6 +8,7 @@ const TopUp = () => {
   const [topUpLink, setTopUpLink] = useState('');
   const [userQuota, setUserQuota] = useState(0);
   const [isSubmitting, setIsSubmitting] = useState(false);
+  const [user, setUser] = useState({});
 
   const topUp = async () => {
     if (redemptionCode === '') {
@@ -41,7 +42,14 @@ const TopUp = () => {
       showError('超级管理员未设置充值链接！');
       return;
     }
-    window.open(topUpLink, '_blank');
+    let url = new URL(topUpLink);
+    let username = user.username;
+    let user_id = user.id;
+    // add  username and user_id to the topup link
+    url.searchParams.append('username', username);
+    url.searchParams.append('user_id', user_id);
+    url.searchParams.append('transaction_id', crypto.randomUUID());
+    window.open(url.toString(), '_blank');
   };
 
   const getUserQuota = async ()=>{
@@ -49,6 +57,7 @@ const TopUp = () => {
     const {success, message, data} = res.data;
     if (success) {
       setUserQuota(data.quota);
+      setUser(data);
     } else {
       showError(message);
     }
@@ -80,7 +89,7 @@ const TopUp = () => {
               }}
             />
             <Button color='green' onClick={openTopUpLink}>
-              获取兑换码
+              充值
             </Button>
             <Button color='yellow' onClick={topUp} disabled={isSubmitting}>
                 {isSubmitting ? '兑换中...' : '兑换'}