feat: modify frontend

Migration 0003 failed when rerank_models table already exists from create_all().
Add table existence check to prevent duplicate creation error in CI environments with cached database.

.github/workflows/build-release-artifacts.yaml

@@ -43,10 +43,10 @@ jobs:
         run: |
           cd /tmp/langbot_build_web/web
           npm install
-          npm run build
+          npx vite build
       - name: Package Output
         run: |
-          cp -r /tmp/langbot_build_web/web/out ./web
+          cp -r /tmp/langbot_build_web/web/dist ./web
       - name: Upload Artifact
         uses: actions/upload-artifact@v4
         with:

.github/workflows/check-i18n.yml

@@ -0,0 +1,25 @@
+name: Check i18n Keys
+
+on:
+  push:
+    branches:
+      - main
+      - master
+
+jobs:
+  check-i18n:
+    name: Check i18n Key Consistency
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: '20'
+
+      - name: Check i18n keys against en-US reference
+        run: node web/scripts/check-i18n.mjs

.github/workflows/publish-to-pypi.yml

@@ -29,8 +29,8 @@ jobs:
           npm install -g pnpm
           pnpm install
           pnpm build
-          mkdir -p ../src/langbot/web/out
-          cp -r out ../src/langbot/web/
+          mkdir -p ../src/langbot/web/dist
+          cp -r dist ../src/langbot/web/
 
       - name: Install the latest version of uv
         uses: astral-sh/setup-uv@v6

.github/workflows/test-migrations.yml

@@ -0,0 +1,171 @@
+name: Test Migrations
+
+on:
+  push:
+    branches:
+      - main
+      - master
+      - dev
+    paths:
+      - 'src/langbot/pkg/persistence/**'
+      - 'src/langbot/pkg/entity/persistence/**'
+  pull_request:
+    types: [opened, synchronize, reopened, ready_for_review]
+    paths:
+      - 'src/langbot/pkg/persistence/**'
+      - 'src/langbot/pkg/entity/persistence/**'
+
+jobs:
+  test-migrations-sqlite:
+    name: Migrations (SQLite)
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+
+      - name: Install dependencies
+        run: uv sync --dev
+
+      - name: Test Alembic upgrade (SQLite)
+        run: |
+          uv run python -c "
+          import asyncio
+          from sqlalchemy.ext.asyncio import create_async_engine
+          from langbot.pkg.entity.persistence.base import Base
+          from langbot.pkg.persistence.alembic_runner import run_alembic_upgrade, run_alembic_stamp, get_alembic_current
+
+          async def main():
+              engine = create_async_engine('sqlite+aiosqlite:///test_migrations.db')
+
+              # Create all tables (simulates existing DB)
+              async with engine.begin() as conn:
+                  await conn.run_sync(Base.metadata.create_all)
+
+              # Stamp baseline
+              await run_alembic_stamp(engine, '0001_baseline')
+              rev = await get_alembic_current(engine)
+              assert rev == '0001_baseline', f'Expected 0001_baseline, got {rev}'
+              print(f'Stamped: {rev}')
+
+              # Upgrade to head
+              await run_alembic_upgrade(engine, 'head')
+              rev = await get_alembic_current(engine)
+              print(f'After upgrade: {rev}')
+              assert rev is not None, 'Expected a revision after upgrade'
+
+              # Verify idempotent
+              await run_alembic_upgrade(engine, 'head')
+              rev2 = await get_alembic_current(engine)
+              assert rev2 == rev, f'Expected {rev}, got {rev2}'
+              print(f'Idempotent check passed: {rev2}')
+
+              # Fresh DB: upgrade from scratch
+              engine2 = create_async_engine('sqlite+aiosqlite:///test_migrations_fresh.db')
+              async with engine2.begin() as conn:
+                  await conn.run_sync(Base.metadata.create_all)
+              await run_alembic_upgrade(engine2, 'head')
+              rev3 = await get_alembic_current(engine2)
+              print(f'Fresh DB upgrade: {rev3}')
+              assert rev3 is not None
+
+              print('All SQLite migration tests passed!')
+
+          asyncio.run(main())
+          "
+
+  test-migrations-postgres:
+    name: Migrations (PostgreSQL)
+    runs-on: ubuntu-latest
+    services:
+      postgres:
+        image: postgres:16
+        env:
+          POSTGRES_USER: langbot
+          POSTGRES_PASSWORD: langbot
+          POSTGRES_DB: langbot_test
+        ports:
+          - 5432:5432
+        options: >-
+          --health-cmd="pg_isready -U langbot"
+          --health-interval=5s
+          --health-timeout=5s
+          --health-retries=5
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+
+      - name: Install dependencies
+        run: uv sync --dev
+
+      - name: Test Alembic upgrade (PostgreSQL)
+        run: |
+          uv run python -c "
+          import asyncio
+          from sqlalchemy.ext.asyncio import create_async_engine
+          from langbot.pkg.entity.persistence.base import Base
+          from langbot.pkg.persistence.alembic_runner import run_alembic_upgrade, run_alembic_stamp, get_alembic_current
+
+          DB_URL = 'postgresql+asyncpg://langbot:langbot@localhost:5432/langbot_test'
+
+          async def main():
+              engine = create_async_engine(DB_URL)
+
+              # Create all tables
+              async with engine.begin() as conn:
+                  await conn.run_sync(Base.metadata.create_all)
+
+              # Stamp baseline
+              await run_alembic_stamp(engine, '0001_baseline')
+              rev = await get_alembic_current(engine)
+              assert rev == '0001_baseline', f'Expected 0001_baseline, got {rev}'
+              print(f'Stamped: {rev}')
+
+              # Upgrade to head
+              await run_alembic_upgrade(engine, 'head')
+              rev = await get_alembic_current(engine)
+              print(f'After upgrade: {rev}')
+              assert rev is not None
+
+              # Verify idempotent
+              await run_alembic_upgrade(engine, 'head')
+              rev2 = await get_alembic_current(engine)
+              assert rev2 == rev, f'Expected {rev}, got {rev2}'
+              print(f'Idempotent check passed: {rev2}')
+
+              # Fresh DB: drop all and upgrade from scratch
+              engine2 = create_async_engine(DB_URL.replace('langbot_test', 'langbot_fresh'))
+
+              # Create fresh database
+              from sqlalchemy import text
+              async with engine.connect() as conn:
+                  await conn.execute(text('COMMIT'))
+                  await conn.execute(text('CREATE DATABASE langbot_fresh'))
+
+              async with engine2.begin() as conn:
+                  await conn.run_sync(Base.metadata.create_all)
+              await run_alembic_upgrade(engine2, 'head')
+              rev3 = await get_alembic_current(engine2)
+              print(f'Fresh DB upgrade: {rev3}')
+              assert rev3 is not None
+
+              print('All PostgreSQL migration tests passed!')
+
+          asyncio.run(main())
+          "

.gitignore

@@ -52,3 +52,6 @@ src/langbot/web/
 /dist
 /build
 *.egg-info
+
+# Next.js build cache (legacy)
+web/.next/

AGENTS.md

@@ -70,7 +70,7 @@ Plugin Runtime automatically starts each installed plugin and interacts through
     - type: must be a specific type, such as feat (new feature), fix (bug fix), docs (documentation), style (code style), refactor (refactoring), perf (performance optimization), etc.
     - scope: the scope of the commit, such as the package name, the file name, the function name, the class name, the module name, etc.
     - subject: the subject of the commit, such as the description of the commit, the reason for the commit, the impact of the commit, etc.
-- If you changed the definition of database entities, please update the migration file in `src/langbot/pkg/persistence/migrations/` and update the constants.py file in `src/langbot/pkg/utils/constants.py` with the new migration number.
+- LangBot uses [Alembic](https://alembic.sqlalchemy.org/) to manage database migrations, supporting both SQLite and PostgreSQL. Migration files are located in `src/langbot/pkg/persistence/alembic/versions/`. If you changed the definition of database entities (ORM models), generate a new migration script by running `uv run python -m langbot.pkg.persistence.alembic_runner autogenerate "description of your change"` in the project root (requires `data/config.yaml` to exist). Review and edit the generated script before committing. Migrations are executed automatically on LangBot startup. For data migrations (e.g. modifying JSON field content), you need to manually add the migration code in the generated script.
 
 ## Some Principles
 

Dockerfile

@@ -4,7 +4,7 @@ WORKDIR /app
 
 COPY web ./web
 
-RUN cd web && npm install && npm run build
+RUN cd web && npm install && npx vite build
 
 FROM python:3.12.7-slim
 
@@ -12,7 +12,7 @@ WORKDIR /app
 
 COPY . .
 
-COPY --from=node /app/web/out ./web/out
+COPY --from=node /app/web/dist ./web/dist
 
 RUN apt update \
     && apt install gcc -y \

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "langbot"
-version = "4.9.5"
+version = "4.9.6"
 description = "Production-grade platform for building agentic IM bots"
 readme = "README.md"
 license-files = ["LICENSE"]
@@ -8,7 +8,7 @@ requires-python = ">=3.11,<4.0"
 dependencies = [
     "aiocqhttp>=1.4.4",
     "aiofiles>=24.1.0",
-    "aiohttp>=3.11.18",
+    "aiohttp>=3.13.4",
     "aioshutil>=1.5",
     "aiosqlite>=0.21.0",
     "anthropic>=0.51.0",
@@ -16,7 +16,7 @@ dependencies = [
     "async-lru>=2.0.5",
     "certifi>=2025.4.26",
     "colorlog~=6.6.0",
-    "cryptography>=44.0.3",
+    "cryptography>=46.0.7",
     "dashscope>=1.25.10",
     "dingtalk-stream>=0.24.0",
     "discord-py>=2.5.2",
@@ -27,7 +27,7 @@ dependencies = [
     "nakuru-project-idk>=0.0.2.1",
     "ollama>=0.4.8",
     "openai>1.0.0",
-    "pillow>=11.2.1",
+    "pillow>=12.2.0",
     "psutil>=7.0.0",
     "pycryptodome>=3.22.0",
     "pydantic>2.0",
@@ -39,6 +39,7 @@ dependencies = [
     "quart-cors>=0.8.0",
     "requests>=2.32.3",
     "slack-sdk>=3.35.0",
+    "alembic>=1.15.0",
     "sqlalchemy[asyncio]>=2.0.40",
     "sqlmodel>=0.0.24",
     "telegramify-markdown>=0.5.1",
@@ -49,7 +50,7 @@ dependencies = [
     "pip>=25.1.1",
     "ruff>=0.11.9",
     "pre-commit>=4.2.0",
-    "uv>=0.7.11",
+    "uv>=0.11.6",
     "mypy>=1.16.0",
     "PyPDF2>=3.0.1",
     "python-docx>=1.1.0",
@@ -60,11 +61,15 @@ dependencies = [
     "ebooklib>=0.18",
     "html2text>=2024.2.26",
     "langchain>=0.2.0",
-    "langchain-text-splitters>=0.0.1",
+    "langchain-core>=1.2.28",
+    "langsmith>=0.7.31",
+    "python-multipart>=0.0.26",
+    "Mako>=1.3.11",
+    "langchain-text-splitters>=1.1.2",
     "chromadb>=1.0.0,<2.0.0",
     "qdrant-client (>=1.15.1,<2.0.0)",
     "pyseekdb==1.1.0.post3",
-    "langbot-plugin==0.3.6",
+    "langbot-plugin==0.3.8",
     "asyncpg>=0.30.0",
     "line-bot-sdk>=3.19.0",
     "tboxsdk>=0.0.10",
@@ -111,12 +116,12 @@ requires = ["setuptools>=61.0", "wheel"]
 build-backend = "setuptools.build_meta"
 
 [tool.setuptools]
-package-data = { "langbot" = ["templates/**", "pkg/provider/modelmgr/requesters/*", "pkg/platform/sources/*", "web/out/**"] }
+package-data = { "langbot" = ["templates/**", "pkg/provider/modelmgr/requesters/*", "pkg/platform/sources/*", "web/dist/**", "pkg/persistence/alembic/**"] }
 
 [dependency-groups]
 dev = [
     "pre-commit>=4.2.0",
-    "pytest>=8.4.1",
+    "pytest>=9.0.3",
     "pytest-asyncio>=1.0.0",
     "pytest-cov>=7.0.0",
     "ruff>=0.11.9",

src/langbot/__init__.py

@@ -1,3 +1,3 @@
 """LangBot - Production-grade platform for building agentic IM bots"""
 
-__version__ = '4.9.5'
+__version__ = '4.9.6'

src/langbot/libs/dingtalk_api/api.py

@@ -182,6 +182,88 @@ class DingTalkClient:
             for handler in self._message_handlers[msg_type]:
                 await handler(event)
 
+    async def _parse_quoted_message(self, replied_msg: dict) -> dict:
+        """Parse the quoted/replied message and extract its content.
+
+        Args:
+            replied_msg: The repliedMsg object from DingTalk message
+
+        Returns:
+            A dict containing the quoted message info with keys:
+            - message_id: The original message ID
+            - msg_type: The message type (text, file, picture, audio, etc.)
+            - content: The text content (if any)
+            - file_url: The file download URL (if file type)
+            - file_name: The file name (if file type)
+            - picture: The picture base64 (if picture type)
+            - audio: The audio base64 (if audio type)
+        """
+        quote_info = {
+            'message_id': replied_msg.get('msgId', ''),
+            'msg_type': replied_msg.get('msgType', ''),
+            'sender_id': replied_msg.get('senderId', ''),
+        }
+
+        msg_type = replied_msg.get('msgType', '')
+        content = replied_msg.get('content', {})
+
+        # Handle content as string (JSON) or dict
+        if isinstance(content, str):
+            try:
+                content = json.loads(content)
+            except (json.JSONDecodeError, TypeError):
+                content = {}
+
+        if msg_type == 'text':
+            # Text message
+            if isinstance(content, dict):
+                quote_info['content'] = content.get('content', '')
+            else:
+                quote_info['content'] = str(content)
+
+        elif msg_type == 'file':
+            # File message
+            download_code = content.get('downloadCode')
+            file_name = content.get('fileName')
+            if download_code and file_name:
+                try:
+                    quote_info['file_url'] = await self.get_file_url(download_code)
+                    quote_info['file_name'] = file_name
+                except Exception as e:
+                    if self.logger:
+                        await self.logger.error(f'Failed to get quoted file URL: {e}')
+
+        elif msg_type == 'picture':
+            # Picture message
+            download_code = content.get('downloadCode')
+            if download_code:
+                try:
+                    quote_info['picture'] = await self.download_image(download_code)
+                except Exception as e:
+                    if self.logger:
+                        await self.logger.error(f'Failed to download quoted image: {e}')
+
+        elif msg_type == 'audio':
+            # Audio message
+            download_code = content.get('downloadCode')
+            if download_code:
+                try:
+                    quote_info['audio'] = await self.get_audio_url(download_code)
+                except Exception as e:
+                    if self.logger:
+                        await self.logger.error(f'Failed to get quoted audio: {e}')
+
+        elif msg_type == 'richText':
+            # Rich text message - extract text content
+            rich_text = content.get('richText', [])
+            texts = []
+            for item in rich_text:
+                if 'text' in item and item['text'] != '\n':
+                    texts.append(item['text'])
+            quote_info['content'] = '\n'.join(texts)
+
+        return quote_info
+
     async def get_message(self, incoming_message: dingtalk_stream.chatbot.ChatbotMessage):
         try:
             # print(json.dumps(incoming_message.to_dict(), indent=4, ensure_ascii=False))
@@ -193,6 +275,15 @@ class DingTalkClient:
             elif str(incoming_message.conversation_type) == '2':
                 message_data['conversation_type'] = 'GroupMessage'
 
+            # Check for quoted/replied message
+            raw_data = incoming_message.to_dict()
+            text_data = raw_data.get('text', {})
+            if isinstance(text_data, dict) and text_data.get('isReplyMsg'):
+                replied_msg = text_data.get('repliedMsg', {})
+                if replied_msg:
+                    quote_info = await self._parse_quoted_message(replied_msg)
+                    message_data['QuotedMessage'] = quote_info
+
             if incoming_message.message_type == 'richText':
                 data = incoming_message.rich_text_content.to_dict()
 
@@ -268,7 +359,25 @@ class DingTalkClient:
 
                 message_data['Type'] = 'image'
             elif incoming_message.message_type == 'audio':
-                message_data['Audio'] = await self.get_audio_url(incoming_message.to_dict()['content']['downloadCode'])
+                raw_content = incoming_message.to_dict().get('content', {})
+                # 兼容处理：如果 content 仍为 JSON 字符串则进行解析
+                if isinstance(raw_content, str):
+                    try:
+                        raw_content = json.loads(raw_content)
+                    except (json.JSONDecodeError, TypeError):
+                        raw_content = {}
+
+                if self.logger:
+                    await self.logger.info(f'DingTalk audio raw content: {json.dumps(raw_content, ensure_ascii=False)}')
+
+                # 提取钉钉自带的语音转写文字（Powered by Qwen）
+                recognition = raw_content.get('recognition', '')
+                if recognition:
+                    message_data['Content'] = recognition
+
+                download_code = raw_content.get('downloadCode')
+                if download_code:
+                    message_data['Audio'] = await self.get_audio_url(download_code)
 
                 message_data['Type'] = 'audio'
             elif incoming_message.message_type == 'file':

src/langbot/libs/dingtalk_api/dingtalkevent.py

@@ -47,6 +47,22 @@ class DingTalkEvent(dict):
     def conversation(self):
         return self.get('conversation_type', '')
 
+    @property
+    def quoted_message(self) -> Optional[Dict[str, Any]]:
+        """Get the quoted/replied message info if this is a reply message.
+
+        Returns:
+            A dict containing:
+            - message_id: The original message ID
+            - msg_type: The message type (text, file, picture, audio, etc.)
+            - content: The text content (if any)
+            - file_url: The file download URL (if file type)
+            - file_name: The file name (if file type)
+            - picture: The picture base64 (if picture type)
+            - audio: The audio base64 (if audio type)
+        """
+        return self.get('QuotedMessage')
+
     def __getattr__(self, key: str) -> Optional[Any]:
         """
         允许通过属性访问数据中的任意字段。

src/langbot/libs/wecom_ai_bot_api/api.py

@@ -71,6 +71,11 @@ class StreamSession:
 class StreamSessionManager:
     """管理 stream 会话的生命周期，并负责队列的生产消费。"""
 
+    # Sessions with registered feedback_ids use a longer TTL to survive the
+    # full like → cancel → dislike feedback flow. Must align with the adapter's
+    # _stream_to_monitoring_msg TTL (wecombot.py).
+    _FEEDBACK_SESSION_TTL = 600  # 10 minutes
+
     def __init__(self, logger: EventLogger, ttl: int = 60) -> None:
         self.logger = logger
 
@@ -214,11 +219,17 @@ class StreamSessionManager:
             session.last_access = time.time()
 
     def cleanup(self) -> None:
-        """定期清理过期会话，防止队列与映射无上限累积。"""
+        """定期清理过期会话，防止队列与映射无上限累积。
+
+        已注册 feedback_id 的会话使用更长的 TTL，确保用户在点赞/取消/点踩流程中
+        不会因为 session 被提前清除而丢失上下文信息。
+        """
         now = time.time()
         expired: list[str] = []
         for stream_id, session in self._sessions.items():
-            if now - session.last_access > self.ttl:
+            # Sessions with registered feedback_ids use a longer TTL
+            effective_ttl = self._FEEDBACK_SESSION_TTL if session.feedback_id else self.ttl
+            if now - session.last_access > effective_ttl:
                 expired.append(stream_id)
 
         for stream_id in expired:
@@ -228,6 +239,9 @@ class StreamSessionManager:
             msg_id = session.msg_id
             if msg_id and self._msg_index.get(msg_id) == stream_id:
                 self._msg_index.pop(msg_id, None)
+            # Clean up feedback index for expired sessions
+            if session.feedback_id:
+                self._feedback_index.pop(session.feedback_id, None)
 
 
 def _decrypt_file(encrypted_data: bytes, aes_key_str: str) -> bytes:
@@ -434,10 +448,10 @@ async def parse_wecom_bot_message(
         }
         if voice_info.get('content'):
             message_data['content'] = voice_info.get('content')
-        if (message_data['voice'].get('filesize') or 0) <= max_inline_file_size:
-            voice_base64 = await _safe_download_as_data_uri(download_url, per_msg_aeskey)
-            if voice_base64:
-                message_data['voice']['base64'] = voice_base64
+        # if (message_data['voice'].get('filesize') or 0) <= max_inline_file_size:
+        #     voice_base64 = await _safe_download_as_data_uri(download_url, per_msg_aeskey)
+        #     if voice_base64:
+        #         message_data['voice']['base64'] = voice_base64
     elif msg_type == 'video':
         video_info = msg_json.get('video', {}) or {}
         download_url = video_info.get('url')
@@ -449,10 +463,12 @@ async def parse_wecom_bot_message(
             'md5sum': video_info.get('md5sum') or video_info.get('md5'),
             'filename': video_info.get('filename') or video_info.get('name'),
         }
-        if (video_data.get('filesize') or 0) <= max_inline_file_size:
-            video_base64 = await _safe_download_as_data_uri(download_url, per_msg_aeskey)
-            if video_base64:
-                video_data['base64'] = video_base64
+        # if (video_data.get('filesize') or 0) <= max_inline_file_size:
+        #     video_base64 = await _safe_download_as_data_uri(download_url, per_msg_aeskey)
+        #     if video_base64:
+        #         video_data['base64'] = video_base64
+        # 应为需要解密，但是目前暂时不能下载到内部进行解密，所以先将下载链接拼接aeskey返回给用户，由插件去处理该链接的下载和解密逻辑
+        video_data['download_url'] = download_url + f'?aeskey={per_msg_aeskey}'
         message_data['video'] = video_data
     elif msg_type == 'file':
         file_info = msg_json.get('file', {}) or {}
@@ -466,12 +482,15 @@ async def parse_wecom_bot_message(
             'download_url': download_url,
             'extra': file_info,
         }
-        if (file_data.get('filesize') or 0) <= max_inline_file_size:
-            file_bytes, dl_filename = await _safe_download(download_url, per_msg_aeskey)
-            if file_bytes:
-                file_data['base64'] = _bytes_to_data_uri(file_bytes)
-                if dl_filename and not file_data.get('filename'):
-                    file_data['filename'] = dl_filename
+        # if (file_data.get('filesize') or 0) <= max_inline_file_size:
+        #     file_bytes, dl_filename = await _safe_download(download_url, per_msg_aeskey)
+        #     if file_bytes:
+        #         file_data['base64'] = _bytes_to_data_uri(file_bytes)
+        #         if dl_filename and not file_data.get('filename'):
+        #             file_data['filename'] = dl_filename
+
+        # 应为需要解密，但是目前暂时不能下载到内部进行解密，所以先将下载链接拼接aeskey返回给用户，由插件去处理该链接的下载和解密逻辑
+        file_data['download_url'] = download_url + f'?aeskey={per_msg_aeskey}'
         message_data['file'] = file_data
     elif msg_type == 'link':
         message_data['link'] = msg_json.get('link', {})
@@ -587,6 +606,120 @@ async def parse_wecom_bot_message(
     if msg_json.get('aibotid'):
         message_data['aibotid'] = msg_json.get('aibotid', '')
 
+    # Handle quote (referenced message) - important for group chat file references
+    quote_info = msg_json.get('quote')
+    if quote_info:
+        quote_data: dict[str, Any] = {}
+        quote_type = quote_info.get('msgtype', '')
+        quote_data['msgtype'] = quote_type
+
+        if quote_type == 'text':
+            quote_data['content'] = quote_info.get('text', {}).get('content', '')
+        elif quote_type == 'image':
+            img_info = quote_info.get('image', {})
+            img_url = img_info.get('url', '')
+            img_aeskey = img_info.get('aeskey', '')
+            base64_data = await _safe_download_as_data_uri(img_url, img_aeskey)
+            if base64_data:
+                quote_data['picurl'] = base64_data
+                quote_data['images'] = [base64_data]
+        elif quote_type == 'file':
+            file_info = quote_info.get('file', {}) or {}
+            download_url = file_info.get('url') or file_info.get('fileurl')
+            item_aeskey = file_info.get('aeskey', '')
+            file_data = {
+                'filename': file_info.get('filename') or file_info.get('name'),
+                'filesize': file_info.get('filesize') or file_info.get('size'),
+                'md5sum': file_info.get('md5sum') or file_info.get('md5'),
+                'sdkfileid': file_info.get('sdkfileid') or file_info.get('fileid'),
+                'download_url': download_url,
+                'extra': file_info,
+            }
+            # Same as private chat: append aeskey to download_url for plugin processing
+            if download_url and item_aeskey:
+                file_data['download_url'] = download_url + f'?aeskey={item_aeskey}'
+            quote_data['file'] = file_data
+        elif quote_type == 'voice':
+            voice_info = quote_info.get('voice', {}) or {}
+            download_url = voice_info.get('url')
+            item_aeskey = voice_info.get('aeskey', '')
+            voice_data = {
+                'url': download_url,
+                'md5sum': voice_info.get('md5sum') or voice_info.get('md5'),
+                'filesize': voice_info.get('filesize') or voice_info.get('size'),
+                'sdkfileid': voice_info.get('sdkfileid') or voice_info.get('fileid'),
+            }
+            if voice_info.get('content'):
+                quote_data['content'] = voice_info.get('content')
+            # Same as private chat: append aeskey to url for plugin processing
+            if download_url and item_aeskey:
+                voice_data['url'] = download_url + f'?aeskey={item_aeskey}'
+            quote_data['voice'] = voice_data
+        elif quote_type == 'video':
+            video_info = quote_info.get('video', {}) or {}
+            download_url = video_info.get('url')
+            item_aeskey = video_info.get('aeskey', '')
+            video_data = {
+                'url': download_url,
+                'filesize': video_info.get('filesize') or video_info.get('size'),
+                'sdkfileid': video_info.get('sdkfileid') or video_info.get('fileid'),
+                'md5sum': video_info.get('md5sum') or video_info.get('md5'),
+                'filename': video_info.get('filename') or video_info.get('name'),
+            }
+            # Same as private chat: append aeskey to download_url for plugin processing
+            if download_url and item_aeskey:
+                video_data['download_url'] = download_url + f'?aeskey={item_aeskey}'
+            quote_data['video'] = video_data
+        elif quote_type == 'link':
+            quote_data['link'] = quote_info.get('link', {})
+            link = quote_data['link']
+            title = link.get('title', '')
+            desc = link.get('description') or link.get('digest', '')
+            quote_data['content'] = '\n'.join(filter(None, [title, desc]))
+        elif quote_type == 'mixed':
+            # Handle mixed type in quote (text + images + files etc.)
+            items = quote_info.get('mixed', {}).get('msg_item', [])
+            texts = []
+            images = []
+            files = []
+            for item in items:
+                item_type = item.get('msgtype')
+                if item_type == 'text':
+                    texts.append(item.get('text', {}).get('content', ''))
+                elif item_type == 'image':
+                    img_info = item.get('image', {})
+                    img_url = img_info.get('url')
+                    img_aeskey = img_info.get('aeskey', '')
+                    base64_data = await _safe_download_as_data_uri(img_url, img_aeskey)
+                    if base64_data:
+                        images.append(base64_data)
+                elif item_type == 'file':
+                    file_info = item.get('file', {}) or {}
+                    download_url = file_info.get('url') or file_info.get('fileurl')
+                    item_aeskey = file_info.get('aeskey', '')
+                    file_data = {
+                        'filename': file_info.get('filename') or file_info.get('name'),
+                        'filesize': file_info.get('filesize') or file_info.get('size'),
+                        'md5sum': file_info.get('md5sum') or file_info.get('md5'),
+                        'sdkfileid': file_info.get('sdkfileid') or file_info.get('fileid'),
+                        'download_url': download_url,
+                        'extra': file_info,
+                    }
+                    # Same as private chat: append aeskey to download_url for plugin processing
+                    if download_url and item_aeskey:
+                        file_data['download_url'] = download_url + f'?aeskey={item_aeskey}'
+                    files.append(file_data)
+            if texts:
+                quote_data['content'] = ' '.join(texts)
+            if images:
+                quote_data['images'] = images
+                quote_data['picurl'] = images[0]
+            if files:
+                quote_data['files'] = files
+                quote_data['file'] = files[0]
+
+        message_data['quote'] = quote_data
+
     return message_data
 
 
@@ -898,35 +1031,38 @@ class WecomBotClient:
             )
 
             session = self.stream_sessions.get_session_by_feedback_id(feedback_id)
+
             if session:
                 await self.logger.info(
                     f'反馈关联到会话: stream_id={session.stream_id}, msg_id={session.msg_id}, user_id={session.user_id}'
                 )
-                for handler in self._message_handlers.get('feedback', []):
-                    try:
-                        await handler(
-                            feedback_id=feedback_id,
-                            feedback_type=feedback_type,
-                            feedback_content=feedback_content,
-                            inaccurate_reasons=inaccurate_reasons,
-                            session=session,
-                        )
-                    except Exception:
-                        await self.logger.error(traceback.format_exc())
-
-                if self._feedback_callback:
-                    try:
-                        await self._feedback_callback(
-                            feedback_id=feedback_id,
-                            feedback_type=feedback_type,
-                            feedback_content=feedback_content,
-                            inaccurate_reasons=inaccurate_reasons,
-                            session=session,
-                        )
-                    except Exception:
-                        await self.logger.error(traceback.format_exc())
             else:
-                await self.logger.warning(f'未找到 feedback_id={feedback_id} 对应的会话')
+                await self.logger.warning(f'未找到 feedback_id={feedback_id} 对应的会话，仍将记录反馈')
+
+            # Dispatch feedback event regardless of session availability
+            for handler in self._message_handlers.get('feedback', []):
+                try:
+                    await handler(
+                        feedback_id=feedback_id,
+                        feedback_type=feedback_type,
+                        feedback_content=feedback_content,
+                        inaccurate_reasons=inaccurate_reasons,
+                        session=session,
+                    )
+                except Exception:
+                    await self.logger.error(traceback.format_exc())
+
+            if self._feedback_callback:
+                try:
+                    await self._feedback_callback(
+                        feedback_id=feedback_id,
+                        feedback_type=feedback_type,
+                        feedback_content=feedback_content,
+                        inaccurate_reasons=inaccurate_reasons,
+                        session=session,
+                    )
+                except Exception:
+                    await self.logger.error(traceback.format_exc())
 
         except Exception:
             await self.logger.error(traceback.format_exc())

src/langbot/libs/wecom_ai_bot_api/wecombotevent.py

@@ -147,3 +147,10 @@ class WecomBotEvent(dict):
         流式消息 ID
         """
         return self.get('stream_id', '')
+
+    @property
+    def quote(self):
+        """
+        引用消息信息（群聊中用户引用其他消息时返回）
+        """
+        return self.get('quote', {})

src/langbot/libs/wecom_ai_bot_api/ws_client.py

@@ -20,7 +20,7 @@ from typing import Any, Callable, Optional
 import aiohttp
 
 from langbot.libs.wecom_ai_bot_api import wecombotevent
-from langbot.libs.wecom_ai_bot_api.api import parse_wecom_bot_message
+from langbot.libs.wecom_ai_bot_api.api import parse_wecom_bot_message, StreamSession
 from langbot.pkg.platform.logger import EventLogger
 
 DEFAULT_WS_URL = 'wss://openws.work.weixin.qq.com'
@@ -96,6 +96,12 @@ class WecomBotWsClient:
         self._stream_ids: dict[str, str] = {}  # msg_id -> req_id|stream_id
         # Dedup: skip sending when content hasn't changed
         self._stream_last_content: dict[str, str] = {}  # msg_id -> last content sent
+        # Stream session info for feedback tracking
+        self._stream_sessions: dict[str, dict] = {}  # msg_id -> session info
+        # Feedback tracking: feedback_id -> session info
+        self._feedback_sessions: dict[str, dict] = {}  # feedback_id -> {msg_id, user_id, chat_id, stream_id, req_id}
+        # msg_id -> feedback_id (for associating feedback with message)
+        self._msg_feedback_ids: dict[str, str] = {}  # msg_id -> feedback_id
 
     # ── Public API ──────────────────────────────────────────────────
 
@@ -164,12 +170,27 @@ class WecomBotWsClient:
 
         return decorator
 
+    def on_feedback(self) -> Callable:
+        """Decorator to register a feedback event handler.
+
+        Same interface as WecomBotClient.on_feedback for compatibility.
+        """
+
+        def decorator(func: Callable):
+            if 'feedback' not in self._message_handlers:
+                self._message_handlers['feedback'] = []
+            self._message_handlers['feedback'].append(func)
+            return func
+
+        return decorator
+
     async def reply_stream(
         self,
         req_id: str,
         stream_id: str,
         content: str,
         finish: bool = False,
+        feedback_id: str = '',
     ) -> Optional[dict]:
         """Send a streaming reply frame.
 
@@ -178,17 +199,22 @@ class WecomBotWsClient:
             stream_id: The stream ID for this streaming session.
             content: The content to send (supports Markdown).
             finish: Whether this is the final chunk.
+            feedback_id: Optional feedback ID for receiving user feedback (like/dislike).
 
         Returns:
             The ACK frame dict, or None on failure.
         """
+        stream_payload = {
+            'id': stream_id,
+            'finish': finish,
+            'content': content,
+        }
+        if feedback_id:
+            stream_payload['feedback'] = {'id': feedback_id}
+
         body = {
             'msgtype': 'stream',
-            'stream': {
-                'id': stream_id,
-                'finish': finish,
-                'content': content,
-            },
+            'stream': stream_payload,
         }
         return await self._send_reply(req_id, body)
 
@@ -253,11 +279,23 @@ class WecomBotWsClient:
             # Skip sending if content hasn't changed (e.g. during tool call argument streaming)
             if not is_final and content == self._stream_last_content.get(msg_id):
                 return True
-            await self.reply_stream(req_id, stream_id, content, finish=is_final)
+
+            # Generate feedback_id for final chunk
+            feedback_id = ''
+            if is_final:
+                feedback_id = _generate_req_id('feedback')
+                self._msg_feedback_ids[msg_id] = feedback_id
+                # Store session info for feedback tracking
+                session_info = self._stream_sessions.get(msg_id)
+                if session_info:
+                    self._feedback_sessions[feedback_id] = session_info
+
+            await self.reply_stream(req_id, stream_id, content, finish=is_final, feedback_id=feedback_id)
             self._stream_last_content[msg_id] = content
             if is_final:
                 self._stream_ids.pop(msg_id, None)
                 self._stream_last_content.pop(msg_id, None)
+                self._stream_sessions.pop(msg_id, None)
             return True
         except Exception:
             await self.logger.error(f'Failed to push stream chunk: {traceback.format_exc()}')
@@ -445,6 +483,15 @@ class WecomBotWsClient:
             msg_id = message_data.get('msgid', '')
             if msg_id:
                 self._stream_ids[msg_id] = f'{req_id}|{stream_id}'
+                # Store session info for feedback tracking
+                self._stream_sessions[msg_id] = {
+                    'req_id': req_id,
+                    'stream_id': stream_id,
+                    'msg_id': msg_id,
+                    'user_id': message_data.get('userid', ''),
+                    'chat_id': message_data.get('chatid', ''),
+                    'chat_type': message_data.get('type', 'single'),
+                }
             message_data['stream_id'] = stream_id
             message_data['req_id'] = req_id
 
@@ -454,7 +501,7 @@ class WecomBotWsClient:
             await self.logger.error(f'Error in message callback: {traceback.format_exc()}')
 
     async def _handle_event_callback(self, frame: dict):
-        """Handle an incoming event callback frame (enter_chat, template_card_event, etc.)."""
+        """Handle an incoming event callback frame (enter_chat, template_card_event, feedback_event, disconnected_event)."""
         try:
             body = frame.get('body', {})
             req_id = frame.get('headers', {}).get('req_id', '')
@@ -479,14 +526,54 @@ class WecomBotWsClient:
             if body.get('chatid'):
                 message_data['chatid'] = body.get('chatid', '')
 
+            if event_type == 'feedback_event':
+                feedback_event = event_info.get('feedback_event', {})
+                feedback_id = feedback_event.get('id', '')
+                feedback_type = feedback_event.get('type', 0)
+                feedback_content = feedback_event.get('content', '')
+                inaccurate_reasons = feedback_event.get('inaccurate_reason_list', [])
+
+                await self.logger.info(
+                    f'收到用户反馈事件: feedback_id={feedback_id}, type={feedback_type}, '
+                    f'content={feedback_content}, reasons={inaccurate_reasons}'
+                )
+
+                # Look up session by feedback_id
+                session_info = self._feedback_sessions.get(feedback_id)
+                session = None
+                if session_info:
+                    session = StreamSession(
+                        stream_id=session_info.get('stream_id', ''),
+                        msg_id=session_info.get('msg_id', ''),
+                        chat_id=session_info.get('chat_id') or None,
+                        user_id=session_info.get('user_id') or None,
+                        feedback_id=feedback_id,
+                    )
+                    await self.logger.info(
+                        f'反馈关联到会话: stream_id={session.stream_id}, msg_id={session.msg_id}, user_id={session.user_id}'
+                    )
+                else:
+                    await self.logger.warning(f'未找到 feedback_id={feedback_id} 对应的会话')
+
+                for handler in self._message_handlers.get('feedback', []):
+                    try:
+                        await handler(
+                            feedback_id=feedback_id,
+                            feedback_type=feedback_type,
+                            feedback_content=feedback_content,
+                            inaccurate_reasons=inaccurate_reasons,
+                            session=session,
+                        )
+                    except Exception:
+                        await self.logger.error(f'Error in feedback handler: {traceback.format_exc()}')
+                return
+
             event = wecombotevent.WecomBotEvent(message_data)
 
-            # Dispatch to event-specific handlers
             if event_type in self._message_handlers:
                 for handler in self._message_handlers[event_type]:
                     await handler(event)
 
-            # Also dispatch to generic 'event' handlers
             if 'event' in self._message_handlers:
                 for handler in self._message_handlers['event']:
                     await handler(event)

src/langbot/pkg/api/http/controller/groups/provider/models.py

@@ -97,3 +97,51 @@ class EmbeddingModelsRouterGroup(group.RouterGroup):
             await self.ap.embedding_models_service.test_embedding_model(model_uuid, json_data)
 
             return self.success()
+
+
+@group.group_class('models/rerank', '/api/v1/provider/models/rerank')
+class RerankModelsRouterGroup(group.RouterGroup):
+    async def initialize(self) -> None:
+        @self.route('', methods=['GET', 'POST'], auth_type=group.AuthType.USER_TOKEN_OR_API_KEY)
+        async def _() -> str:
+            if quart.request.method == 'GET':
+                provider_uuid = quart.request.args.get('provider_uuid')
+                if provider_uuid:
+                    return self.success(
+                        data={
+                            'models': await self.ap.rerank_models_service.get_rerank_models_by_provider(provider_uuid)
+                        }
+                    )
+                return self.success(data={'models': await self.ap.rerank_models_service.get_rerank_models()})
+            elif quart.request.method == 'POST':
+                json_data = await quart.request.json
+                model_uuid = await self.ap.rerank_models_service.create_rerank_model(json_data)
+                return self.success(data={'uuid': model_uuid})
+
+        @self.route('/<model_uuid>', methods=['GET', 'PUT', 'DELETE'], auth_type=group.AuthType.USER_TOKEN_OR_API_KEY)
+        async def _(model_uuid: str) -> str:
+            if quart.request.method == 'GET':
+                model = await self.ap.rerank_models_service.get_rerank_model(model_uuid)
+
+                if model is None:
+                    return self.http_status(404, -1, 'model not found')
+
+                return self.success(data={'model': model})
+            elif quart.request.method == 'PUT':
+                json_data = await quart.request.json
+
+                await self.ap.rerank_models_service.update_rerank_model(model_uuid, json_data)
+
+                return self.success()
+            elif quart.request.method == 'DELETE':
+                await self.ap.rerank_models_service.delete_rerank_model(model_uuid)
+
+                return self.success()
+
+        @self.route('/<model_uuid>/test', methods=['POST'], auth_type=group.AuthType.USER_TOKEN_OR_API_KEY)
+        async def _(model_uuid: str) -> str:
+            json_data = await quart.request.json
+
+            await self.ap.rerank_models_service.test_rerank_model(model_uuid, json_data)
+
+            return self.success()

src/langbot/pkg/api/http/controller/groups/provider/providers.py

@@ -15,6 +15,7 @@ class ModelProvidersRouterGroup(group.RouterGroup):
                     counts = await self.ap.provider_service.get_provider_model_counts(provider['uuid'])
                     provider['llm_count'] = counts['llm_count']
                     provider['embedding_count'] = counts['embedding_count']
+                    provider['rerank_count'] = counts['rerank_count']
                 return self.success(data={'providers': providers})
             elif quart.request.method == 'POST':
                 json_data = await quart.request.json
@@ -32,6 +33,7 @@ class ModelProvidersRouterGroup(group.RouterGroup):
                 counts = await self.ap.provider_service.get_provider_model_counts(provider_uuid)
                 provider['llm_count'] = counts['llm_count']
                 provider['embedding_count'] = counts['embedding_count']
+                provider['rerank_count'] = counts['rerank_count']
                 return self.success(data={'provider': provider})
             elif quart.request.method == 'PUT':
                 json_data = await quart.request.json
@@ -43,3 +45,12 @@ class ModelProvidersRouterGroup(group.RouterGroup):
                     return self.success()
                 except ValueError as e:
                     return self.http_status(400, -1, str(e))
+
+        @self.route('/<provider_uuid>/scan-models', methods=['GET'], auth_type=group.AuthType.USER_TOKEN_OR_API_KEY)
+        async def _(provider_uuid: str) -> str:
+            try:
+                model_type = quart.request.args.get('type')
+                result = await self.ap.provider_service.scan_provider_models(provider_uuid, model_type)
+                return self.success(data=result)
+            except ValueError as e:
+                return self.http_status(400, -1, str(e))

src/langbot/pkg/api/http/controller/groups/resources/tools.py

@@ -0,0 +1,45 @@
+from __future__ import annotations
+
+from ... import group
+
+
+@group.group_class('tools', '/api/v1/tools')
+class ToolsRouterGroup(group.RouterGroup):
+    async def initialize(self) -> None:
+        @self.route('', methods=['GET'], auth_type=group.AuthType.USER_TOKEN)
+        async def _() -> str:
+            """获取所有可用工具列表"""
+            tools = await self.ap.tool_mgr.get_all_tools()
+
+            tool_list = []
+            for tool in tools:
+                tool_list.append(
+                    {
+                        'name': tool.name,
+                        'description': tool.description,
+                        'human_desc': tool.human_desc,
+                        'parameters': tool.parameters,
+                    }
+                )
+
+            return self.success(data={'tools': tool_list})
+
+        @self.route('/<tool_name>', methods=['GET'], auth_type=group.AuthType.USER_TOKEN)
+        async def _(tool_name: str) -> str:
+            """获取特定工具详情"""
+            tools = await self.ap.tool_mgr.get_all_tools()
+
+            for tool in tools:
+                if tool.name == tool_name:
+                    return self.success(
+                        data={
+                            'tool': {
+                                'name': tool.name,
+                                'description': tool.description,
+                                'human_desc': tool.human_desc,
+                                'parameters': tool.parameters,
+                            }
+                        }
+                    )
+
+            return self.http_status(404, -1, f'Tool not found: {tool_name}')

src/langbot/pkg/api/http/controller/main.py

@@ -105,23 +105,24 @@ class HTTPController:
             ):
                 if os.path.exists(os.path.join(frontend_path, path + '.html')):
                     path += '.html'
-                elif path.startswith('home/'):
-                    # SPA fallback for /home/* sub-routes.
-                    # Entity detail views use query params (e.g. /home/bots?id=uuid),
-                    # so the pre-rendered list page is served directly via path + '.html'.
-                    # This fallback handles any remaining unmatched sub-paths.
-                    segments = path.rstrip('/').split('/')
+                elif not path.startswith('api/'):
+                    # SPA fallback: serve index.html for all non-API, non-static routes
+                    # so that React Router can handle client-side routing (Vite SPA).
+                    # For /home/* sub-routes, first try parent .html files (pre-rendered pages).
+                    if path.startswith('home/'):
+                        segments = path.rstrip('/').split('/')
+                        for i in range(len(segments) - 1, 0, -1):
+                            parent_path = '/'.join(segments[:i]) + '.html'
+                            if os.path.exists(os.path.join(frontend_path, parent_path)):
+                                response = await quart.send_from_directory(
+                                    frontend_path, parent_path, mimetype='text/html'
+                                )
+                                response.headers['Cache-Control'] = 'no-cache, no-store, must-revalidate'
+                                response.headers['Pragma'] = 'no-cache'
+                                response.headers['Expires'] = '0'
+                                return response
 
-                    # Walk up parent segments looking for matching .html files
-                    for i in range(len(segments) - 1, 0, -1):
-                        parent_path = '/'.join(segments[:i]) + '.html'
-                        if os.path.exists(os.path.join(frontend_path, parent_path)):
-                            response = await quart.send_from_directory(frontend_path, parent_path, mimetype='text/html')
-                            response.headers['Cache-Control'] = 'no-cache, no-store, must-revalidate'
-                            response.headers['Pragma'] = 'no-cache'
-                            response.headers['Expires'] = '0'
-                            return response
-                    # Final fallback to index.html for /home/* routes
+                    # Fallback to index.html for SPA client-side routing
                     response = await quart.send_from_directory(frontend_path, 'index.html', mimetype='text/html')
                     response.headers['Cache-Control'] = 'no-cache, no-store, must-revalidate'
                     response.headers['Pragma'] = 'no-cache'

src/langbot/pkg/api/http/service/model.py

@@ -367,3 +367,162 @@ class EmbeddingModelsService:
             input_text=['Hello, world!'],
             extra_args={},
         )
+
+
+class RerankModelsService:
+    ap: app.Application
+
+    def __init__(self, ap: app.Application) -> None:
+        self.ap = ap
+
+    async def get_rerank_models(self) -> list[dict]:
+        """Get all rerank models with provider info"""
+        result = await self.ap.persistence_mgr.execute_async(sqlalchemy.select(persistence_model.RerankModel))
+        models = result.all()
+
+        providers_result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(persistence_model.ModelProvider)
+        )
+        providers = {p.uuid: p for p in providers_result.all()}
+
+        models_list = []
+        for model in models:
+            model_dict = self.ap.persistence_mgr.serialize_model(persistence_model.RerankModel, model)
+            provider = providers.get(model.provider_uuid)
+            if provider:
+                provider_dict = self.ap.persistence_mgr.serialize_model(persistence_model.ModelProvider, provider)
+                model_dict['provider'] = _parse_provider_api_keys(provider_dict)
+            models_list.append(model_dict)
+
+        return models_list
+
+    async def get_rerank_models_by_provider(self, provider_uuid: str) -> list[dict]:
+        """Get rerank models by provider UUID"""
+        result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(persistence_model.RerankModel).where(
+                persistence_model.RerankModel.provider_uuid == provider_uuid
+            )
+        )
+        models = result.all()
+        return [self.ap.persistence_mgr.serialize_model(persistence_model.RerankModel, m) for m in models]
+
+    async def create_rerank_model(self, model_data: dict, preserve_uuid: bool = False) -> str:
+        """Create a new rerank model"""
+        if not preserve_uuid:
+            model_data['uuid'] = str(uuid.uuid4())
+
+        if 'provider' in model_data:
+            provider_data = model_data.pop('provider')
+            if provider_data.get('uuid'):
+                model_data['provider_uuid'] = provider_data['uuid']
+            else:
+                provider_uuid = await self.ap.provider_service.find_or_create_provider(
+                    requester=provider_data.get('requester', ''),
+                    base_url=provider_data.get('base_url', ''),
+                    api_keys=provider_data.get('api_keys', []),
+                )
+                model_data['provider_uuid'] = provider_uuid
+
+        await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.insert(persistence_model.RerankModel).values(**model_data)
+        )
+
+        runtime_provider = self.ap.model_mgr.provider_dict.get(model_data['provider_uuid'])
+        if runtime_provider is None:
+            raise Exception('provider not found')
+
+        runtime_rerank_model = await self.ap.model_mgr.load_rerank_model_with_provider(
+            persistence_model.RerankModel(**model_data),
+            runtime_provider,
+        )
+        self.ap.model_mgr.rerank_models.append(runtime_rerank_model)
+
+        return model_data['uuid']
+
+    async def get_rerank_model(self, model_uuid: str) -> dict | None:
+        """Get a single rerank model with provider info"""
+        result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(persistence_model.RerankModel).where(persistence_model.RerankModel.uuid == model_uuid)
+        )
+        model = result.first()
+        if model is None:
+            return None
+
+        model_dict = self.ap.persistence_mgr.serialize_model(persistence_model.RerankModel, model)
+
+        provider_result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(persistence_model.ModelProvider).where(
+                persistence_model.ModelProvider.uuid == model.provider_uuid
+            )
+        )
+        provider = provider_result.first()
+        if provider:
+            provider_dict = self.ap.persistence_mgr.serialize_model(persistence_model.ModelProvider, provider)
+            model_dict['provider'] = _parse_provider_api_keys(provider_dict)
+
+        return model_dict
+
+    async def update_rerank_model(self, model_uuid: str, model_data: dict) -> None:
+        """Update an existing rerank model"""
+        if 'uuid' in model_data:
+            del model_data['uuid']
+
+        if 'provider' in model_data:
+            provider_data = model_data.pop('provider')
+            if provider_data.get('uuid'):
+                model_data['provider_uuid'] = provider_data['uuid']
+            else:
+                provider_uuid = await self.ap.provider_service.find_or_create_provider(
+                    requester=provider_data.get('requester', ''),
+                    base_url=provider_data.get('base_url', ''),
+                    api_keys=provider_data.get('api_keys', []),
+                )
+                model_data['provider_uuid'] = provider_uuid
+
+        await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.update(persistence_model.RerankModel)
+            .where(persistence_model.RerankModel.uuid == model_uuid)
+            .values(**model_data)
+        )
+
+        await self.ap.model_mgr.remove_rerank_model(model_uuid)
+
+        runtime_provider = self.ap.model_mgr.provider_dict.get(model_data['provider_uuid'])
+        if runtime_provider is None:
+            raise Exception('provider not found')
+
+        runtime_rerank_model = await self.ap.model_mgr.load_rerank_model_with_provider(
+            persistence_model.RerankModel(**model_data),
+            runtime_provider,
+        )
+        self.ap.model_mgr.rerank_models.append(runtime_rerank_model)
+
+    async def delete_rerank_model(self, model_uuid: str) -> None:
+        """Delete a rerank model"""
+        await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.delete(persistence_model.RerankModel).where(persistence_model.RerankModel.uuid == model_uuid)
+        )
+        await self.ap.model_mgr.remove_rerank_model(model_uuid)
+
+    async def test_rerank_model(self, model_uuid: str, model_data: dict) -> None:
+        """Test a rerank model"""
+        runtime_rerank_model: model_requester.RuntimeRerankModel | None = None
+
+        if model_uuid != '_':
+            for model in self.ap.model_mgr.rerank_models:
+                if model.model_entity.uuid == model_uuid:
+                    runtime_rerank_model = model
+                    break
+            if runtime_rerank_model is None:
+                raise Exception('model not found')
+        else:
+            runtime_rerank_model = await self.ap.model_mgr.init_temporary_runtime_rerank_model(model_data)
+
+        await runtime_rerank_model.provider.invoke_rerank(
+            model=runtime_rerank_model,
+            query='What is artificial intelligence?',
+            documents=[
+                'Artificial intelligence is a branch of computer science.',
+                'The weather is nice today.',
+            ],
+        )

src/langbot/pkg/api/http/service/monitoring.py

@@ -1224,30 +1224,83 @@ class MonitoringService:
         """
         import json
 
-        record_id = str(uuid.uuid4())
-        record_data = {
-            'id': record_id,
-            'timestamp': datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None),
-            'feedback_id': feedback_id,
-            'feedback_type': feedback_type,
-            'feedback_content': feedback_content,
-            'inaccurate_reasons': json.dumps(inaccurate_reasons, ensure_ascii=False) if inaccurate_reasons else None,
-            'bot_id': bot_id,
-            'bot_name': bot_name,
-            'pipeline_id': pipeline_id,
-            'pipeline_name': pipeline_name,
-            'session_id': session_id,
-            'message_id': message_id,
-            'stream_id': stream_id,
-            'user_id': user_id,
-            'platform': platform,
-        }
+        now = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
+        reasons_json = json.dumps(inaccurate_reasons, ensure_ascii=False) if inaccurate_reasons else None
 
-        await self.ap.persistence_mgr.execute_async(
-            sqlalchemy.insert(persistence_monitoring.MonitoringFeedback).values(record_data)
+        MonitoringFeedback = persistence_monitoring.MonitoringFeedback
+
+        # Handle cancel feedback (type=3): delete existing record
+        if feedback_type == 3:
+            await self.ap.persistence_mgr.execute_async(
+                sqlalchemy.delete(MonitoringFeedback).where(MonitoringFeedback.feedback_id == feedback_id)
+            )
+            return None
+
+        # Check if record with this feedback_id already exists
+        existing_result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(MonitoringFeedback).where(MonitoringFeedback.feedback_id == feedback_id)
         )
+        existing_row = existing_result.first()
 
-        return record_id
+        if existing_row:
+            # UPDATE existing record
+            existing = existing_row[0] if isinstance(existing_row, tuple) else existing_row
+            await self.ap.persistence_mgr.execute_async(
+                sqlalchemy.update(MonitoringFeedback)
+                .where(MonitoringFeedback.feedback_id == feedback_id)
+                .values(
+                    timestamp=now,
+                    feedback_type=feedback_type,
+                    feedback_content=feedback_content,
+                    inaccurate_reasons=reasons_json,
+                    bot_id=bot_id or existing.bot_id,
+                    bot_name=bot_name or existing.bot_name,
+                    pipeline_id=pipeline_id or existing.pipeline_id,
+                    pipeline_name=pipeline_name or existing.pipeline_name,
+                    session_id=session_id or existing.session_id,
+                    message_id=message_id or existing.message_id,
+                    stream_id=stream_id or existing.stream_id,
+                    user_id=user_id or existing.user_id,
+                    platform=platform or existing.platform,
+                )
+            )
+            return existing.id
+        else:
+            # INSERT new record with IntegrityError defense
+            record_id = str(uuid.uuid4())
+            record_data = {
+                'id': record_id,
+                'timestamp': now,
+                'feedback_id': feedback_id,
+                'feedback_type': feedback_type,
+                'feedback_content': feedback_content,
+                'inaccurate_reasons': reasons_json,
+                'bot_id': bot_id,
+                'bot_name': bot_name,
+                'pipeline_id': pipeline_id,
+                'pipeline_name': pipeline_name,
+                'session_id': session_id,
+                'message_id': message_id,
+                'stream_id': stream_id,
+                'user_id': user_id,
+                'platform': platform,
+            }
+            try:
+                await self.ap.persistence_mgr.execute_async(sqlalchemy.insert(MonitoringFeedback).values(record_data))
+                return record_id
+            except Exception:
+                # UNIQUE constraint conflict (concurrent feedback for same feedback_id)
+                await self.ap.persistence_mgr.execute_async(
+                    sqlalchemy.update(MonitoringFeedback)
+                    .where(MonitoringFeedback.feedback_id == feedback_id)
+                    .values(
+                        timestamp=now,
+                        feedback_type=feedback_type,
+                        feedback_content=feedback_content,
+                        inaccurate_reasons=reasons_json,
+                    )
+                )
+                return feedback_id
 
     async def get_feedback_stats(
         self,

src/langbot/pkg/api/http/service/provider.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import uuid
+import traceback
 
 import sqlalchemy
 
@@ -97,6 +98,14 @@ class ModelProviderService:
         if embedding_result.first() is not None:
             raise ValueError('Cannot delete provider: Embedding models still reference it')
 
+        rerank_result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(persistence_model.RerankModel).where(
+                persistence_model.RerankModel.provider_uuid == provider_uuid
+            )
+        )
+        if rerank_result.first() is not None:
+            raise ValueError('Cannot delete provider: Rerank models still reference it')
+
         await self.ap.persistence_mgr.execute_async(
             sqlalchemy.delete(persistence_model.ModelProvider).where(
                 persistence_model.ModelProvider.uuid == provider_uuid
@@ -121,7 +130,14 @@ class ModelProviderService:
         )
         embedding_count = embedding_result.scalar() or 0
 
-        return {'llm_count': llm_count, 'embedding_count': embedding_count}
+        rerank_result = await self.ap.persistence_mgr.execute_async(
+            sqlalchemy.select(sqlalchemy.func.count())
+            .select_from(persistence_model.RerankModel)
+            .where(persistence_model.RerankModel.provider_uuid == provider_uuid)
+        )
+        rerank_count = rerank_result.scalar() or 0
+
+        return {'llm_count': llm_count, 'embedding_count': embedding_count, 'rerank_count': rerank_count}
 
     async def find_or_create_provider(self, requester: str, base_url: str, api_keys: list) -> str:
         """Find existing provider or create new one"""
@@ -164,3 +180,66 @@ class ModelProviderService:
             .values(api_keys=[api_key])
         )
         await self.ap.model_mgr.reload_provider('00000000-0000-0000-0000-000000000000')
+
+    async def scan_provider_models(self, provider_uuid: str, model_type: str | None = None) -> dict:
+        provider = await self.get_provider(provider_uuid)
+        if provider is None:
+            raise ValueError('provider not found')
+
+        runtime_provider = await self.ap.model_mgr.load_provider(provider)
+
+        try:
+            scan_result = await runtime_provider.requester.scan_models(
+                runtime_provider.token_mgr.get_token() if runtime_provider.token_mgr.tokens else None
+            )
+        except NotImplementedError:
+            raise ValueError('current provider does not support model scanning')
+        except Exception as exc:
+            self.ap.logger.warning(
+                f'Failed to scan models for provider {provider_uuid}: {exc}\n{traceback.format_exc()}'
+            )
+            raise ValueError(str(exc)) from exc
+
+        if isinstance(scan_result, dict):
+            scanned_models = scan_result.get('models', [])
+            debug_info = scan_result.get('debug')
+        else:
+            scanned_models = scan_result
+            debug_info = None
+
+        llm_models = await self.ap.llm_model_service.get_llm_models_by_provider(provider_uuid)
+        embedding_models = await self.ap.embedding_models_service.get_embedding_models_by_provider(provider_uuid)
+        existing_llm_names = {model['name'] for model in llm_models}
+        existing_embedding_names = {model['name'] for model in embedding_models}
+
+        filtered_models = []
+        for model in scanned_models:
+            scanned_type = model.get('type', 'llm')
+            if model_type and scanned_type != model_type:
+                continue
+
+            model_name = model.get('name') or model.get('id')
+            if not model_name:
+                continue
+
+            filtered_models.append(
+                {
+                    'id': model.get('id', model_name),
+                    'name': model_name,
+                    'type': scanned_type,
+                    'abilities': model.get('abilities', []),
+                    'display_name': model.get('display_name'),
+                    'description': model.get('description'),
+                    'context_length': model.get('context_length'),
+                    'owned_by': model.get('owned_by'),
+                    'input_modalities': model.get('input_modalities', []),
+                    'output_modalities': model.get('output_modalities', []),
+                    'already_added': (
+                        model_name in existing_embedding_names
+                        if scanned_type == 'embedding'
+                        else model_name in existing_llm_names
+                    ),
+                }
+            )
+
+        return {'models': filtered_models, 'debug': debug_info}

src/langbot/pkg/api/http/service/user.py

@@ -65,8 +65,8 @@ class UserService:
 
         user_obj = result_list[0]
 
-        # Check if this is a Space account
-        if user_obj.account_type == 'space':
+        # Check if this user has a local password set
+        if not user_obj.password:
             raise ValueError('请使用 Space 账户登录')
 
         ph = argon2.PasswordHasher()
@@ -108,9 +108,8 @@ class UserService:
         if user_obj is None:
             raise ValueError('User not found')
 
-        # Space accounts cannot change password locally
-        if user_obj.account_type == 'space':
-            raise ValueError('Space account cannot change password locally')
+        if not user_obj.password:
+            raise ValueError('No local password set, please set a password first')
 
         ph.verify(user_obj.password, current_password)
 

src/langbot/pkg/core/app.py

@@ -133,6 +133,8 @@ class Application:
 
     embedding_models_service: model_service.EmbeddingModelsService = None
 
+    rerank_models_service: model_service.RerankModelsService = None
+
     provider_service: provider_service.ModelProviderService = None
 
     pipeline_service: pipeline_service.PipelineService = None

src/langbot/pkg/core/stages/build_app.py

@@ -61,6 +61,9 @@ class BuildAppStage(stage.BootingStage):
         embedding_models_service_inst = model_service.EmbeddingModelsService(ap)
         ap.embedding_models_service = embedding_models_service_inst
 
+        rerank_models_service_inst = model_service.RerankModelsService(ap)
+        ap.rerank_models_service = rerank_models_service_inst
+
         provider_service_inst = provider_service.ModelProviderService(ap)
         ap.provider_service = provider_service_inst
 

src/langbot/pkg/core/stages/load_config.py

@@ -80,8 +80,12 @@ def _apply_env_overrides_to_config(cfg: dict) -> dict:
             if i == len(keys) - 1:
                 # At the final key
                 if key in current:
-                    if isinstance(current[key], (dict, list)):
-                        # Skip dict and list types
+                    if isinstance(current[key], list):
+                        # Convert comma-separated string to list
+                        # e.g., SYSTEM__DISABLED_ADAPTERS="aiocqhttp,dingtalk"
+                        current[key] = [item.strip() for item in env_value.split(',') if item.strip()]
+                    elif isinstance(current[key], dict):
+                        # Skip dict types
                         pass
                     else:
                         # Valid scalar value - convert and set it

src/langbot/pkg/entity/persistence/bot.py

@@ -16,6 +16,7 @@ class Bot(Base):
     enable = sqlalchemy.Column(sqlalchemy.Boolean, nullable=False, default=False)
     use_pipeline_name = sqlalchemy.Column(sqlalchemy.String(255), nullable=True)
     use_pipeline_uuid = sqlalchemy.Column(sqlalchemy.String(255), nullable=True)
+    pipeline_routing_rules = sqlalchemy.Column(sqlalchemy.JSON, nullable=False, server_default='[]')
     created_at = sqlalchemy.Column(sqlalchemy.DateTime, nullable=False, server_default=sqlalchemy.func.now())
     updated_at = sqlalchemy.Column(
         sqlalchemy.DateTime,

src/langbot/pkg/entity/persistence/model.py

@@ -59,3 +59,22 @@ class EmbeddingModel(Base):
         server_default=sqlalchemy.func.now(),
         onupdate=sqlalchemy.func.now(),
     )
+
+
+class RerankModel(Base):
+    """Rerank model"""
+
+    __tablename__ = 'rerank_models'
+
+    uuid = sqlalchemy.Column(sqlalchemy.String(255), primary_key=True, unique=True)
+    name = sqlalchemy.Column(sqlalchemy.String(255), nullable=False)
+    provider_uuid = sqlalchemy.Column(sqlalchemy.String(255), nullable=False)
+    extra_args = sqlalchemy.Column(sqlalchemy.JSON, nullable=False, default={})
+    prefered_ranking = sqlalchemy.Column(sqlalchemy.Integer, nullable=False, default=0)
+    created_at = sqlalchemy.Column(sqlalchemy.DateTime, nullable=False, server_default=sqlalchemy.func.now())
+    updated_at = sqlalchemy.Column(
+        sqlalchemy.DateTime,
+        nullable=False,
+        server_default=sqlalchemy.func.now(),
+        onupdate=sqlalchemy.func.now(),
+    )

src/langbot/pkg/persistence/alembic/env.py

@@ -0,0 +1,51 @@
+"""Alembic environment for LangBot.
+
+This env.py is designed to be called programmatically (not via CLI).
+It supports both SQLite and PostgreSQL.
+
+The sync connection is passed via config attributes by the runner.
+"""
+
+from __future__ import annotations
+
+from alembic import context
+from sqlalchemy.engine import Connection
+
+from langbot.pkg.entity.persistence.base import Base
+
+target_metadata = Base.metadata
+
+
+def run_migrations_offline() -> None:
+    """Run migrations in 'offline' mode — emit SQL without a live connection."""
+    url = context.config.get_main_option('sqlalchemy.url')
+    context.configure(
+        url=url,
+        target_metadata=target_metadata,
+        literal_binds=True,
+        dialect_opts={'paramstyle': 'named'},
+    )
+    with context.begin_transaction():
+        context.run_migrations()
+
+
+def run_migrations_online() -> None:
+    """Run migrations with a live sync connection passed via config attributes."""
+    connection: Connection = context.config.attributes.get('connection')
+    if connection is None:
+        raise RuntimeError('connection not provided in alembic config attributes')
+
+    context.configure(
+        connection=connection,
+        target_metadata=target_metadata,
+        # render_as_batch=True is critical for SQLite ALTER TABLE support
+        render_as_batch=True,
+    )
+    with context.begin_transaction():
+        context.run_migrations()
+
+
+if context.is_offline_mode():
+    run_migrations_offline()
+else:
+    run_migrations_online()

src/langbot/pkg/persistence/alembic/script.py.mako

@@ -0,0 +1,24 @@
+# Alembic script.py.mako — template for auto-generated revisions
+"""${message}
+
+Revision ID: ${up_revision}
+Revises: ${down_revision | comma,n}
+Create Date: ${create_date}
+"""
+from alembic import op
+import sqlalchemy as sa
+${imports if imports else ""}
+
+# revision identifiers
+revision = ${repr(up_revision)}
+down_revision = ${repr(down_revision)}
+branch_labels = ${repr(branch_labels)}
+depends_on = ${repr(depends_on)}
+
+
+def upgrade() -> None:
+    ${upgrades if upgrades else "pass"}
+
+
+def downgrade() -> None:
+    ${downgrades if downgrades else "pass"}

src/langbot/pkg/persistence/alembic/versions/0001_baseline.py

@@ -0,0 +1,24 @@
+"""baseline: stamp existing schema (db version 25)
+
+This is a no-op migration that marks the starting point for Alembic.
+All tables already exist via create_all() + legacy DBMigration system.
+
+Revision ID: 0001_baseline
+Revises: None
+Create Date: 2026-04-08
+"""
+
+revision = '0001_baseline'
+down_revision = None
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # No-op: existing schema is already at database_version=25
+    # This revision serves as the Alembic baseline.
+    pass
+
+
+def downgrade() -> None:
+    pass

src/langbot/pkg/persistence/alembic/versions/0002_sample.py

@@ -0,0 +1,62 @@
+"""example: sample migration demonstrating Alembic patterns
+
+This is a SAMPLE showing how to write migrations that work
+seamlessly across SQLite and PostgreSQL. Delete or adapt as needed.
+
+Revision ID: 0002_sample
+Revises: 0001_baseline
+Create Date: 2026-04-08
+
+Patterns demonstrated:
+  1. Schema change (add column) — works on both DBs via render_as_batch
+  2. Data migration (read + modify JSON) — pure SQLAlchemy, no dialect branching
+"""
+
+revision = '0002_sample'
+down_revision = '0001_baseline'
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    """
+    EXAMPLE: Uncomment to use. This shows the patterns.
+
+    # --- Pattern 1: Schema change (add/drop column) ---
+    # render_as_batch=True in env.py makes this work on SQLite too.
+    #
+    # op.add_column('pipelines', sa.Column('description', sa.String(512), server_default=''))
+
+    # --- Pattern 2: Data migration (read + modify JSON field) ---
+    # No if/else for sqlite vs postgres needed!
+    #
+    # conn = op.get_bind()
+    # rows = conn.execute(sa.text("SELECT uuid, config FROM pipelines")).fetchall()
+    # for row in rows:
+    #     config = json.loads(row[1]) if isinstance(row[1], str) else row[1]
+    #     # Modify the config
+    #     config.setdefault('ai', {}).setdefault('some_new_key', 'default_value')
+    #     conn.execute(
+    #         sa.text("UPDATE pipelines SET config = :cfg WHERE uuid = :uuid"),
+    #         {"cfg": json.dumps(config), "uuid": row[0]}
+    #     )
+
+    # --- Pattern 3: Create a new table ---
+    #
+    # op.create_table(
+    #     'audit_log',
+    #     sa.Column('id', sa.Integer, primary_key=True, autoincrement=True),
+    #     sa.Column('action', sa.String(255), nullable=False),
+    #     sa.Column('detail', sa.Text),
+    #     sa.Column('created_at', sa.DateTime, server_default=sa.func.now()),
+    # )
+    """
+    pass
+
+
+def downgrade() -> None:
+    """
+    # op.drop_column('pipelines', 'description')
+    # op.drop_table('audit_log')
+    """
+    pass

src/langbot/pkg/persistence/alembic/versions/0003_add_rerank_models.py

@@ -0,0 +1,35 @@
+"""add rerank_models table
+
+Revision ID: 0003_add_rerank_models
+Revises: 0002_sample
+Create Date: 2026-04-19
+"""
+
+import sqlalchemy as sa
+from alembic import op
+
+revision = '0003_add_rerank_models'
+down_revision = '0002_sample'
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # Check if table already exists (may have been created by create_all())
+    conn = op.get_bind()
+    inspector = sa.inspect(conn)
+    if 'rerank_models' not in inspector.get_table_names():
+        op.create_table(
+            'rerank_models',
+            sa.Column('uuid', sa.String(255), primary_key=True, unique=True),
+            sa.Column('name', sa.String(255), nullable=False),
+            sa.Column('provider_uuid', sa.String(255), nullable=False),
+            sa.Column('extra_args', sa.JSON, nullable=False, server_default='{}'),
+            sa.Column('prefered_ranking', sa.Integer, nullable=False, server_default='0'),
+            sa.Column('created_at', sa.DateTime, nullable=False, server_default=sa.func.now()),
+            sa.Column('updated_at', sa.DateTime, nullable=False, server_default=sa.func.now()),
+        )
+
+
+def downgrade() -> None:
+    op.drop_table('rerank_models')

src/langbot/pkg/persistence/alembic_runner.py

@@ -0,0 +1,150 @@
+"""Programmatic Alembic runner for LangBot.
+
+Usage from async code:
+    from langbot.pkg.persistence.alembic_runner import run_alembic_upgrade
+    await run_alembic_upgrade(async_engine)
+
+CLI usage (autogenerate):
+    python -m langbot.pkg.persistence.alembic_runner autogenerate "add description column"
+    python -m langbot.pkg.persistence.alembic_runner upgrade
+    python -m langbot.pkg.persistence.alembic_runner current
+"""
+
+from __future__ import annotations
+
+import os
+from typing import TYPE_CHECKING
+
+from alembic.config import Config
+from alembic import command
+from alembic.runtime.migration import MigrationContext
+
+if TYPE_CHECKING:
+    from sqlalchemy.ext.asyncio import AsyncEngine
+    from sqlalchemy.engine import Connection
+
+
+_ALEMBIC_DIR = os.path.join(os.path.dirname(__file__), 'alembic')
+
+
+def _build_config(connection: Connection) -> Config:
+    """Build an Alembic Config with sync connection attached."""
+    cfg = Config()
+    cfg.set_main_option('script_location', _ALEMBIC_DIR)
+    cfg.attributes['connection'] = connection
+    return cfg
+
+
+def _do_upgrade(connection: Connection, revision: str = 'head') -> None:
+    """Synchronous upgrade — runs inside run_sync."""
+    cfg = _build_config(connection)
+    command.upgrade(cfg, revision)
+
+
+def _do_stamp(connection: Connection, revision: str = 'head') -> None:
+    """Synchronous stamp — runs inside run_sync."""
+    cfg = _build_config(connection)
+    command.stamp(cfg, revision)
+
+
+def _do_get_current(connection: Connection) -> str | None:
+    """Get current alembic revision synchronously."""
+    ctx = MigrationContext.configure(connection)
+    return ctx.get_current_revision()
+
+
+def _do_autogenerate(connection: Connection, message: str = 'auto migration') -> None:
+    """Synchronous autogenerate — runs inside run_sync."""
+    cfg = _build_config(connection)
+    command.revision(cfg, message=message, autogenerate=True)
+
+
+async def run_alembic_upgrade(async_engine: AsyncEngine, revision: str = 'head') -> None:
+    """Run Alembic upgrade to the given revision."""
+    async with async_engine.connect() as conn:
+        await conn.run_sync(_do_upgrade, revision)
+        await conn.commit()
+
+
+async def run_alembic_stamp(async_engine: AsyncEngine, revision: str = 'head') -> None:
+    """Stamp the database with a revision without running migrations."""
+    async with async_engine.connect() as conn:
+        await conn.run_sync(_do_stamp, revision)
+        await conn.commit()
+
+
+async def get_alembic_current(async_engine: AsyncEngine) -> str | None:
+    """Get current alembic revision, or None if not stamped."""
+    async with async_engine.connect() as conn:
+        return await conn.run_sync(_do_get_current)
+
+
+async def run_alembic_autogenerate(async_engine: AsyncEngine, message: str = 'auto migration') -> None:
+    """Compare ORM models against DB schema and generate a migration script."""
+    async with async_engine.connect() as conn:
+        await conn.run_sync(_do_autogenerate, message)
+
+
+# CLI entrypoint: python -m langbot.pkg.persistence.alembic_runner <command> [args]
+if __name__ == '__main__':
+    import sys
+    import asyncio
+
+    def _get_engine():
+        """Create engine from data/config.yaml or default SQLite."""
+        from sqlalchemy.ext.asyncio import create_async_engine
+
+        try:
+            import yaml
+
+            with open('data/config.yaml') as f:
+                config = yaml.safe_load(f)
+            db_cfg = config.get('database', {})
+            db_type = db_cfg.get('use', 'sqlite')
+            if db_type == 'postgresql':
+                pg = db_cfg.get('postgresql', {})
+                url = (
+                    f'postgresql+asyncpg://{pg.get("user", "postgres")}:{pg.get("password", "postgres")}'
+                    f'@{pg.get("host", "127.0.0.1")}:{pg.get("port", 5432)}/{pg.get("database", "postgres")}'
+                )
+            else:
+                path = db_cfg.get('sqlite', {}).get('path', 'data/langbot.db')
+                url = f'sqlite+aiosqlite:///{path}'
+        except Exception:
+            url = 'sqlite+aiosqlite:///data/langbot.db'
+
+        return create_async_engine(url)
+
+    def main():
+        if len(sys.argv) < 2:
+            print('Usage: python -m langbot.pkg.persistence.alembic_runner <command> [args]')
+            print('Commands:')
+            print('  autogenerate "message"  — Generate migration from ORM model diff')
+            print('  upgrade [revision]      — Upgrade database (default: head)')
+            print('  stamp [revision]        — Stamp revision without running (default: head)')
+            print('  current                 — Show current revision')
+            sys.exit(1)
+
+        cmd = sys.argv[1]
+        engine = _get_engine()
+
+        if cmd == 'autogenerate':
+            msg = sys.argv[2] if len(sys.argv) > 2 else 'auto migration'
+            asyncio.run(run_alembic_autogenerate(engine, msg))
+            print(f'Migration generated: {msg}')
+        elif cmd == 'upgrade':
+            rev = sys.argv[2] if len(sys.argv) > 2 else 'head'
+            asyncio.run(run_alembic_upgrade(engine, rev))
+            print(f'Upgraded to: {rev}')
+        elif cmd == 'stamp':
+            rev = sys.argv[2] if len(sys.argv) > 2 else 'head'
+            asyncio.run(run_alembic_stamp(engine, rev))
+            print(f'Stamped: {rev}')
+        elif cmd == 'current':
+            rev = asyncio.run(get_alembic_current(engine))
+            print(f'Current revision: {rev}')
+        else:
+            print(f'Unknown command: {cmd}')
+            sys.exit(1)
+
+    main()

src/langbot/pkg/persistence/mgr.py

@@ -76,6 +76,9 @@ class PersistenceManager:
 
             self.ap.logger.info(f'Successfully upgraded database to version {last_migration_number}.')
 
+        # Run Alembic migrations (new migration system)
+        await self._run_alembic_migrations()
+
         await self.write_space_model_providers()
 
     async def create_tables(self):
@@ -135,6 +138,28 @@ class PersistenceManager:
 
     # =================================
 
+    async def _run_alembic_migrations(self):
+        """Run Alembic-based migrations after legacy migrations complete."""
+        from . import alembic_runner
+
+        engine = self.get_db_engine()
+
+        try:
+            current_rev = await alembic_runner.get_alembic_current(engine)
+
+            if current_rev is None:
+                # First time: stamp baseline so Alembic knows existing schema is up-to-date
+                self.ap.logger.info('Alembic: no revision found, stamping baseline...')
+                await alembic_runner.run_alembic_stamp(engine, '0001_baseline')
+                current_rev = '0001_baseline'
+
+            # Upgrade to head
+            await alembic_runner.run_alembic_upgrade(engine, 'head')
+            self.ap.logger.info('Alembic migrations completed.')
+        except Exception as e:
+            self.ap.logger.error(f'Alembic migration failed: {e}', exc_info=True)
+            raise
+
     async def execute_async(self, *args, **kwargs) -> sqlalchemy.engine.cursor.CursorResult:
         async with self.get_db_engine().connect() as conn:
             result = await conn.execute(*args, **kwargs)

src/langbot/pkg/persistence/migrations/dbm025_bot_pipeline_routing_rules.py

@@ -0,0 +1,15 @@
+import sqlalchemy
+from .. import migration
+
+
+@migration.migration_class(25)
+class DBMigrateBotPipelineRoutingRules(migration.DBMigration):
+    """Add pipeline_routing_rules column to bots table"""
+
+    async def upgrade(self):
+        sql_text = sqlalchemy.text("ALTER TABLE bots ADD COLUMN pipeline_routing_rules JSON NOT NULL DEFAULT '[]'")
+        await self.ap.persistence_mgr.execute_async(sql_text)
+
+    async def downgrade(self):
+        sql_text = sqlalchemy.text('ALTER TABLE bots DROP COLUMN pipeline_routing_rules')
+        await self.ap.persistence_mgr.execute_async(sql_text)

src/langbot/pkg/pipeline/aggregator.py

@@ -37,6 +37,7 @@ class PendingMessage:
     message_chain: platform_message.MessageChain
     adapter: abstract_platform_adapter.AbstractMessagePlatformAdapter
     pipeline_uuid: typing.Optional[str]
+    routed_by_rule: bool = False
     timestamp: float = field(default_factory=time.time)
 
 
@@ -125,6 +126,7 @@ class MessageAggregator:
         message_chain: platform_message.MessageChain,
         adapter: abstract_platform_adapter.AbstractMessagePlatformAdapter,
         pipeline_uuid: typing.Optional[str] = None,
+        routed_by_rule: bool = False,
     ) -> None:
         """Add a message to the aggregation buffer
 
@@ -145,6 +147,7 @@ class MessageAggregator:
                 message_chain=message_chain,
                 adapter=adapter,
                 pipeline_uuid=pipeline_uuid,
+                routed_by_rule=routed_by_rule,
             )
             return
 
@@ -159,6 +162,7 @@ class MessageAggregator:
             message_chain=message_chain,
             adapter=adapter,
             pipeline_uuid=pipeline_uuid,
+            routed_by_rule=routed_by_rule,
         )
 
         force_flush = False
@@ -217,6 +221,7 @@ class MessageAggregator:
                 message_chain=msg.message_chain,
                 adapter=msg.adapter,
                 pipeline_uuid=msg.pipeline_uuid,
+                routed_by_rule=msg.routed_by_rule,
             )
             return
 
@@ -231,6 +236,7 @@ class MessageAggregator:
             message_chain=merged_msg.message_chain,
             adapter=merged_msg.adapter,
             pipeline_uuid=merged_msg.pipeline_uuid,
+            routed_by_rule=merged_msg.routed_by_rule,
         )
 
     def _merge_messages(self, messages: list[PendingMessage]) -> PendingMessage:

src/langbot/pkg/pipeline/controller.py

@@ -63,6 +63,14 @@ class Controller:
                                 pipeline = await self.ap.pipeline_mgr.get_pipeline_by_uuid(pipeline_uuid)
                                 if pipeline:
                                     await pipeline.run(selected_query)
+                                else:
+                                    self.ap.logger.warning(
+                                        f'Pipeline {pipeline_uuid} not found for query {selected_query.query_id}, query dropped'
+                                    )
+                            else:
+                                self.ap.logger.warning(
+                                    f'No pipeline_uuid for query {selected_query.query_id}, query dropped'
+                                )
 
                         async with self.ap.query_pool:
                             (await self.ap.sess_mgr.get_session(selected_query))._semaphore.release()

src/langbot/pkg/pipeline/pipelinemgr.py

@@ -297,6 +297,9 @@ class RuntimePipeline:
             )
             # Store message_id in query variables for LLM call monitoring
             query.variables['_monitoring_message_id'] = message_id
+            # Notify adapter so it can map platform-specific IDs to monitoring message ID
+            if hasattr(query.adapter, 'on_monitoring_message_created'):
+                await query.adapter.on_monitoring_message_created(query, message_id)
         except Exception as e:
             self.ap.logger.error(f'Failed to record query start: {e}')
 
@@ -323,6 +326,9 @@ class RuntimePipeline:
             event_ctx = await self.ap.plugin_connector.emit_event(event_obj, bound_plugins)
 
             if event_ctx.is_prevented_default():
+                self.ap.logger.debug(
+                    f'MessageReceived event prevented default for query {query.query_id}, pipeline={pipeline_name}'
+                )
                 return
 
             self.ap.logger.debug(f'Processing query {query.query_id}')

src/langbot/pkg/pipeline/pool.py

@@ -41,6 +41,7 @@ class QueryPool:
         message_chain: platform_message.MessageChain,
         adapter: abstract_platform_adapter.AbstractMessagePlatformAdapter,
         pipeline_uuid: typing.Optional[str] = None,
+        routed_by_rule: bool = False,
     ) -> pipeline_query.Query:
         async with self.condition:
             query_id = self.query_id_counter
@@ -52,7 +53,7 @@ class QueryPool:
                 sender_id=sender_id,
                 message_event=message_event,
                 message_chain=message_chain,
-                variables={},
+                variables={'_routed_by_rule': routed_by_rule},
                 resp_messages=[],
                 resp_message_chain=[],
                 adapter=adapter,

src/langbot/pkg/pipeline/preproc/preproc.py

@@ -160,7 +160,6 @@ class PreProcessor(stage.PipelineStage):
                 elif me.url:
                     content_list.append(provider_message.ContentElement.from_file_url(me.url, 'voice'))
             elif isinstance(me, platform_message.File):
-                # if me.url is not None:
                 content_list.append(provider_message.ContentElement.from_file_url(me.url, me.name))
             elif isinstance(me, platform_message.Quote) and quote_msg:
                 for msg in me.origin:
@@ -172,6 +171,15 @@ class PreProcessor(stage.PipelineStage):
                         ):
                             if msg.base64 is not None:
                                 content_list.append(provider_message.ContentElement.from_image_base64(msg.base64))
+                    elif isinstance(msg, platform_message.File):
+                        content_list.append(provider_message.ContentElement.from_file_url(msg.url, msg.name))
+                    elif isinstance(msg, platform_message.Voice):
+                        if msg.base64:
+                            content_list.append(
+                                provider_message.ContentElement.from_file_base64(msg.base64, 'voice.silk')
+                            )
+                        elif msg.url:
+                            content_list.append(provider_message.ContentElement.from_file_url(msg.url, 'voice'))
 
         query.variables['user_message_text'] = plain_text
 

src/langbot/pkg/pipeline/process/handlers/chat.py

@@ -61,6 +61,9 @@ class ChatMessageHandler(handler.MessageHandler):
 
                 yield entities.StageProcessResult(result_type=entities.ResultType.CONTINUE, new_query=query)
             else:
+                self.ap.logger.debug(
+                    f'NormalMessageReceived event prevented default for query {query.query_id} without reply'
+                )
                 yield entities.StageProcessResult(result_type=entities.ResultType.INTERRUPT, new_query=query)
         else:
             if event_ctx.event.user_message_alter is not None:
@@ -205,6 +208,7 @@ class ChatMessageHandler(handler.MessageHandler):
                         'model_name': model_name,
                         'version': constants.semantic_version,
                         'instance_id': constants.instance_id,
+                        'edition': constants.edition,
                         'pipeline_plugins': pipeline_plugins,
                         'error': locals().get('error_info', None),
                         'timestamp': datetime.utcnow().isoformat(),

src/langbot/pkg/pipeline/resprule/resprule.py

@@ -37,6 +37,10 @@ class GroupRespondRuleCheckStage(stage.PipelineStage):
         if query.launcher_type.value != 'group':  # 只处理群消息
             return entities.StageProcessResult(result_type=entities.ResultType.CONTINUE, new_query=query)
 
+        # 通过路由规则明确指定的流水线，跳过群响应规则检查
+        if query.variables and query.variables.get('_routed_by_rule', False):
+            return entities.StageProcessResult(result_type=entities.ResultType.CONTINUE, new_query=query)
+
         rules = query.pipeline_config['trigger']['group-respond-rules']
 
         use_rule = rules

src/langbot/pkg/platform/botmgr.py

@@ -1,6 +1,8 @@
 from __future__ import annotations
 
 import asyncio
+import json
+import re
 import traceback
 import sqlalchemy
 
@@ -52,6 +54,148 @@ class RuntimeBot:
         self.task_context = taskmgr.TaskContext()
         self.logger = logger
 
+    @staticmethod
+    def _match_operator(actual: str, operator: str, expected: str) -> bool:
+        """Evaluate a single operator condition."""
+        if operator == 'eq':
+            return actual == expected
+        elif operator == 'neq':
+            return actual != expected
+        elif operator == 'contains':
+            return expected in actual
+        elif operator == 'not_contains':
+            return expected not in actual
+        elif operator == 'starts_with':
+            return actual.startswith(expected)
+        elif operator == 'regex':
+            try:
+                return bool(re.search(expected, actual))
+            except re.error:
+                return False
+        return False
+
+    PIPELINE_DISCARD = '__discard__'
+    PIPELINE_DISCARD_DISPLAY_NAME = 'Discarded'
+
+    def resolve_pipeline_uuid(
+        self,
+        launcher_type: str,
+        launcher_id: str,
+        message_text: str,
+        message_element_types: list[str] | None = None,
+    ) -> tuple[str | None, bool]:
+        """Resolve pipeline UUID based on routing rules.
+
+        Rules are evaluated in order; first match wins.
+        Falls back to use_pipeline_uuid if no rule matches.
+
+        Rule types:
+          - launcher_type: session type ("person" / "group")
+          - launcher_id: session / group id
+          - message_content: message text content
+          - message_has_element: message contains element of given type
+            (Image, Voice, File, Forward, Face, At, AtAll, Quote)
+            Operators: eq (has), neq (doesn't have)
+
+        Operators: eq, neq, contains, not_contains, starts_with, regex
+
+        When pipeline_uuid is ``__discard__``, the message should be
+        silently dropped by the caller.
+
+        Returns:
+            tuple: (pipeline_uuid, routed_by_rule) - routed_by_rule is True
+            when a routing rule matched, False when falling back to default.
+        """
+        rules = self.bot_entity.pipeline_routing_rules or []
+        element_type_set = set(message_element_types or [])
+
+        for rule in rules:
+            rule_type = rule.get('type')
+            operator = rule.get('operator', 'eq')
+            rule_value = rule.get('value', '')
+            target_uuid = rule.get('pipeline_uuid')
+            if not rule_type or not target_uuid:
+                continue
+
+            if rule_type == 'launcher_type':
+                if self._match_operator(launcher_type, operator, rule_value):
+                    return target_uuid, True
+            elif rule_type == 'launcher_id':
+                if self._match_operator(str(launcher_id), operator, str(rule_value)):
+                    return target_uuid, True
+            elif rule_type == 'message_content':
+                if self._match_operator(message_text, operator, rule_value):
+                    return target_uuid, True
+            elif rule_type == 'message_has_element':
+                has_element = rule_value in element_type_set
+                if operator == 'eq' and has_element:
+                    return target_uuid, True
+                elif operator == 'neq' and not has_element:
+                    return target_uuid, True
+
+        return self.bot_entity.use_pipeline_uuid, False
+
+    async def _record_discarded_message(
+        self,
+        launcher_type: provider_session.LauncherTypes,
+        launcher_id: str | int,
+        sender_id: str | int,
+        message_event: platform_events.MessageEvent,
+        message_chain: platform_message.MessageChain,
+    ) -> None:
+        """Record a discarded message in the monitoring system."""
+        try:
+            if hasattr(message_chain, 'model_dump'):
+                message_content = json.dumps(message_chain.model_dump(), ensure_ascii=False)
+            else:
+                message_content = str(message_chain)
+
+            sender_name = None
+            if hasattr(message_event, 'sender'):
+                if hasattr(message_event.sender, 'nickname'):
+                    sender_name = message_event.sender.nickname
+                elif hasattr(message_event.sender, 'member_name'):
+                    sender_name = message_event.sender.member_name
+
+            # Use the same session_id format as monitoring_helper.py
+            session_id = f'{launcher_type}_{launcher_id}'
+            platform = launcher_type.value if hasattr(launcher_type, 'value') else str(launcher_type)
+
+            await self.ap.monitoring_service.record_message(
+                bot_id=self.bot_entity.uuid,
+                bot_name=self.bot_entity.name or self.bot_entity.uuid,
+                pipeline_id=self.PIPELINE_DISCARD,
+                pipeline_name=self.PIPELINE_DISCARD_DISPLAY_NAME,
+                message_content=message_content,
+                session_id=session_id,
+                status='discarded',
+                level='info',
+                platform=platform,
+                user_id=str(sender_id),
+                user_name=sender_name,
+            )
+
+            # Ensure the session exists so the message appears in the session monitor.
+            # Don't overwrite pipeline info — a session may have messages from
+            # multiple pipelines; discarding shouldn't change the displayed pipeline.
+            session_updated = await self.ap.monitoring_service.update_session_activity(
+                session_id,
+            )
+            if not session_updated:
+                # No session yet (first message for this launcher was discarded).
+                await self.ap.monitoring_service.record_session_start(
+                    session_id=session_id,
+                    bot_id=self.bot_entity.uuid,
+                    bot_name=self.bot_entity.name or self.bot_entity.uuid,
+                    pipeline_id=self.PIPELINE_DISCARD,
+                    pipeline_name=self.PIPELINE_DISCARD_DISPLAY_NAME,
+                    platform=platform,
+                    user_id=str(sender_id),
+                    user_name=sender_name,
+                )
+        except Exception as e:
+            await self.logger.error(f'Failed to record discarded message: {e}')
+
     async def initialize(self):
         async def on_friend_message(
             event: platform_events.FriendMessage,
@@ -83,6 +227,23 @@ class RuntimeBot:
                     if custom_launcher_id:
                         launcher_id = custom_launcher_id
 
+                message_text = str(event.message_chain)
+                element_types = [comp.type for comp in event.message_chain]
+                pipeline_uuid, routed_by_rule = self.resolve_pipeline_uuid(
+                    'person', launcher_id, message_text, element_types
+                )
+
+                if pipeline_uuid == self.PIPELINE_DISCARD:
+                    await self.logger.info('Person message discarded by routing rule')
+                    await self._record_discarded_message(
+                        provider_session.LauncherTypes.PERSON,
+                        launcher_id,
+                        event.sender.id,
+                        event,
+                        event.message_chain,
+                    )
+                    return
+
                 await self.ap.msg_aggregator.add_message(
                     bot_uuid=self.bot_entity.uuid,
                     launcher_type=provider_session.LauncherTypes.PERSON,
@@ -91,7 +252,8 @@ class RuntimeBot:
                     message_event=event,
                     message_chain=event.message_chain,
                     adapter=adapter,
-                    pipeline_uuid=self.bot_entity.use_pipeline_uuid,
+                    pipeline_uuid=pipeline_uuid,
+                    routed_by_rule=routed_by_rule,
                 )
             else:
                 await self.logger.info('Pipeline skipped for person message due to webhook response')
@@ -126,6 +288,23 @@ class RuntimeBot:
                     if custom_launcher_id:
                         launcher_id = custom_launcher_id
 
+                message_text = str(event.message_chain)
+                element_types = [comp.type for comp in event.message_chain]
+                pipeline_uuid, routed_by_rule = self.resolve_pipeline_uuid(
+                    'group', launcher_id, message_text, element_types
+                )
+
+                if pipeline_uuid == self.PIPELINE_DISCARD:
+                    await self.logger.info('Group message discarded by routing rule')
+                    await self._record_discarded_message(
+                        provider_session.LauncherTypes.GROUP,
+                        launcher_id,
+                        event.sender.id,
+                        event,
+                        event.message_chain,
+                    )
+                    return
+
                 await self.ap.msg_aggregator.add_message(
                     bot_uuid=self.bot_entity.uuid,
                     launcher_type=provider_session.LauncherTypes.GROUP,
@@ -134,7 +313,8 @@ class RuntimeBot:
                     message_event=event,
                     message_chain=event.message_chain,
                     adapter=adapter,
-                    pipeline_uuid=self.bot_entity.use_pipeline_uuid,
+                    pipeline_uuid=pipeline_uuid,
+                    routed_by_rule=routed_by_rule,
                 )
             else:
                 await self.logger.info('Pipeline skipped for group message due to webhook response')
@@ -241,12 +421,20 @@ class PlatformManager:
         # delete all bot log images
         await self.ap.storage_mgr.storage_provider.delete_dir_recursive('bot_log_images')
 
+        disabled_adapters = self.ap.instance_config.data.get('system', {}).get('disabled_adapters', []) or []
+
         self.adapter_components = self.ap.discover.get_components_by_kind('MessagePlatformAdapter')
         adapter_dict: dict[str, type[abstract_platform_adapter.AbstractMessagePlatformAdapter]] = {}
         for component in self.adapter_components:
+            if component.metadata.name in disabled_adapters:
+                continue
             adapter_dict[component.metadata.name] = component.get_python_component_class()
         self.adapter_dict = adapter_dict
 
+        # Filter out disabled adapters from components list (for API responses)
+        if disabled_adapters:
+            self.adapter_components = [c for c in self.adapter_components if c.metadata.name not in disabled_adapters]
+
         # initialize websocket adapter
         websocket_adapter_class = self.adapter_dict['websocket']
         websocket_logger = EventLogger(name='websocket-adapter', ap=self.ap)

src/langbot/pkg/platform/sources/dingtalk.py

@@ -71,7 +71,8 @@ class DingTalkMessageConverter(abstract_platform_adapter.AbstractMessageConverte
                     yiri_msg_list.append(platform_message.Image(base64=element['Picture']))
         else:
             # 回退到原有简单逻辑
-            if event.content:
+            # 对于音频消息，content 来自 recognition 转写文字，在下方音频处理块中统一处理
+            if event.content and event.type != 'audio':
                 text_content = event.content.replace('@' + bot_name, '')
                 yiri_msg_list.append(platform_message.Plain(text=text_content))
             if event.picture:
@@ -81,7 +82,38 @@ class DingTalkMessageConverter(abstract_platform_adapter.AbstractMessageConverte
         if event.file:
             yiri_msg_list.append(platform_message.File(url=event.file, name=event.name))
         if event.audio:
-            yiri_msg_list.append(platform_message.Voice(base64=event.audio))
+            # 优先使用钉钉自带的语音转写文字（recognition字段）
+            if event.content and event.type == 'audio':
+                yiri_msg_list.append(platform_message.Plain(text=event.content))
+            else:
+                yiri_msg_list.append(platform_message.Voice(base64=event.audio))
+
+        # Handle quoted/replied message - extract content as top-level components
+        # so that plugins like FileReader can process them the same way as direct messages
+        if event.quoted_message:
+            quote_info = event.quoted_message
+            msg_type = quote_info.get('msg_type', '')
+
+            # Process quoted file - add as top-level File component (same as private chat)
+            if msg_type == 'file' and quote_info.get('file_url'):
+                file_name = quote_info.get('file_name', 'file')
+                yiri_msg_list.append(platform_message.File(url=quote_info['file_url'], name=file_name))
+
+            # Process quoted image - add as top-level Image component
+            elif msg_type == 'picture' and quote_info.get('picture'):
+                yiri_msg_list.append(platform_message.Image(base64=quote_info['picture']))
+
+            # Process quoted audio - add as top-level Voice component
+            elif msg_type == 'audio' and quote_info.get('audio'):
+                yiri_msg_list.append(platform_message.Voice(base64=quote_info['audio']))
+
+            # Process quoted text - add as Plain text with context prefix
+            elif msg_type == 'text' and quote_info.get('content'):
+                yiri_msg_list.append(platform_message.Plain(text=f'[引用消息] {quote_info["content"]}'))
+
+            # Process quoted rich text - add as Plain text with context prefix
+            elif msg_type == 'richText' and quote_info.get('content'):
+                yiri_msg_list.append(platform_message.Plain(text=f'[引用消息] {quote_info["content"]}'))
 
         chain = platform_message.MessageChain(yiri_msg_list)
 

src/langbot/pkg/platform/sources/lark.py

@@ -709,21 +709,29 @@ class LarkEventConverter(abstract_platform_adapter.AbstractEventConverter):
         message_chain = await LarkMessageConverter.target2yiri(event.event.message, api_client)
 
         # Check for quote/reply message
+        # Extract files/images/voice from quote and add them as top-level components
+        # so that plugins like FileReader can process them the same way as direct messages
         quote_message_id = LarkEventConverter._extract_quote_message_id(event.event.message)
         if quote_message_id:
             quote_chain = await LarkEventConverter._fetch_quoted_message(quote_message_id, api_client)
             if quote_chain:
                 # Filter out Source component from quoted chain, keep only content
-                quote_origin = platform_message.MessageChain(
-                    [comp for comp in quote_chain if not isinstance(comp, platform_message.Source)]
-                )
-                if quote_origin:
-                    message_chain.append(
-                        platform_message.Quote(
-                            message_id=quote_message_id,
-                            origin=quote_origin,
-                        )
-                    )
+                quote_components = [comp for comp in quote_chain if not isinstance(comp, platform_message.Source)]
+
+                # Add quoted content as top-level components instead of wrapping in Quote
+                for comp in quote_components:
+                    if isinstance(comp, platform_message.File):
+                        # Add file as top-level component (same as direct message)
+                        message_chain.append(comp)
+                    elif isinstance(comp, platform_message.Image):
+                        # Add image as top-level component
+                        message_chain.append(comp)
+                    elif isinstance(comp, platform_message.Voice):
+                        # Add voice as top-level component
+                        message_chain.append(comp)
+                    elif isinstance(comp, platform_message.Plain):
+                        # Add text with context prefix
+                        message_chain.append(platform_message.Plain(text=f'[引用消息] {comp.text}'))
 
         if event.event.message.chat_type == 'p2p':
             return platform_events.FriendMessage(
@@ -779,6 +787,13 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
 
     card_id_dict: dict[str, str]  # 消息id到卡片id的映射，便于创建卡片后的发送消息到指定卡片
 
+    # Monitoring message ID mapping for feedback correlation
+    # Temp: user Lark message ID → monitoring_message_id (populated by on_monitoring_message_created, consumed by create_message_card)
+    pending_monitoring_msg: dict[str, str]
+    # Final: reply Lark message ID → (monitoring_message_id, timestamp) (used by feedback callbacks)
+    reply_to_monitoring_msg: dict[str, tuple[str, float]]
+    _MONITORING_MAPPING_TTL = 600  # 10 minutes
+
     seq: int  # 用于在发送卡片消息中识别消息顺序，直接以seq作为标识
     bot_uuid: str = None  # 机器人UUID
     app_ticket: str = None  # 商店应用用到
@@ -797,8 +812,71 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
         def sync_on_message(event: lark_oapi.im.v1.P2ImMessageReceiveV1):
             asyncio.create_task(on_message(event))
 
+        def sync_on_card_action(event):
+            try:
+                action_value_obj = getattr(getattr(event.event, 'action', None), 'value', {})
+                action_value = action_value_obj.get('feedback', '') if isinstance(action_value_obj, dict) else ''
+
+                if action_value == '有帮助':
+                    feedback_type = 1
+                elif action_value == '无帮助':
+                    feedback_type = 2
+                else:
+                    from lark_oapi.event.callback.model.p2_card_action_trigger import P2CardActionTriggerResponse
+
+                    return P2CardActionTriggerResponse({'toast': {'type': 'success', 'content': '操作成功'}})
+
+                operator = getattr(event.event, 'operator', None)
+                context = getattr(event.event, 'context', None)
+
+                user_id = getattr(operator, 'open_id', None) or getattr(operator, 'user_id', None)
+                open_chat_id = getattr(context, 'open_chat_id', None)
+                open_message_id = getattr(context, 'open_message_id', None)
+
+                if open_chat_id:
+                    session_id = f'group_{open_chat_id}'
+                elif user_id:
+                    session_id = f'person_{user_id}'
+                else:
+                    session_id = None
+
+                # Resolve monitoring message ID from reply message mapping
+                monitoring_msg_id = None
+                if open_message_id and open_message_id in self.reply_to_monitoring_msg:
+                    monitoring_msg_id = self.reply_to_monitoring_msg[open_message_id][0]
+
+                feedback_event = platform_events.FeedbackEvent(
+                    feedback_id=getattr(event.header, 'event_id', str(uuid.uuid4())),
+                    feedback_type=feedback_type,
+                    feedback_content=action_value,
+                    user_id=user_id,
+                    session_id=session_id,
+                    message_id=open_message_id,
+                    stream_id=monitoring_msg_id,
+                    source_platform_object=event,
+                )
+
+                if platform_events.FeedbackEvent in self.listeners:
+                    loop = asyncio.get_event_loop()
+                    if loop.is_running():
+                        asyncio.create_task(self.listeners[platform_events.FeedbackEvent](feedback_event, self))
+                    else:
+                        loop.run_until_complete(self.listeners[platform_events.FeedbackEvent](feedback_event, self))
+
+                from lark_oapi.event.callback.model.p2_card_action_trigger import P2CardActionTriggerResponse
+
+                return P2CardActionTriggerResponse({'toast': {'type': 'success', 'content': '感谢您的反馈'}})
+            except Exception:
+                asyncio.create_task(self.logger.error(f'Error in lark card action callback: {traceback.format_exc()}'))
+                from lark_oapi.event.callback.model.p2_card_action_trigger import P2CardActionTriggerResponse
+
+                return P2CardActionTriggerResponse({'toast': {'type': 'error', 'content': '反馈处理失败'}})
+
         event_handler = (
-            lark_oapi.EventDispatcherHandler.builder('', '').register_p2_im_message_receive_v1(sync_on_message).build()
+            lark_oapi.EventDispatcherHandler.builder('', '')
+            .register_p2_im_message_receive_v1(sync_on_message)
+            .register_p2_card_action_trigger(sync_on_card_action)
+            .build()
         )
 
         bot_account_id = config['bot_name']
@@ -813,6 +891,8 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
             logger=logger,
             lark_tenant_key=config.get('lark_tenant_key', ''),
             card_id_dict={},
+            pending_monitoring_msg={},
+            reply_to_monitoring_msg={},
             seq=1,
             listeners={},
             quart_app=quart_app,
@@ -953,6 +1033,22 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
             is_stream = True
         return is_stream
 
+    async def on_monitoring_message_created(self, query, monitoring_message_id: str):
+        """Called by pipeline after monitoring message is created, to map user message ID to monitoring message ID."""
+        try:
+            user_msg_id = query.message_event.message_chain.message_id
+            if user_msg_id:
+                self.pending_monitoring_msg[user_msg_id] = monitoring_message_id
+        except Exception as e:
+            await self.logger.debug(f'Failed to map message to monitoring message: {e}')
+
+    def _cleanup_monitoring_mapping(self):
+        """Remove entries older than TTL from the reply-to-monitoring mapping."""
+        now = time.time()
+        expired = [k for k, (_, ts) in self.reply_to_monitoring_msg.items() if now - ts > self._MONITORING_MAPPING_TTL]
+        for k in expired:
+            del self.reply_to_monitoring_msg[k]
+
     async def create_card_id(self, message_id):
         try:
             # self.logger.debug('飞书支持stream输出,创建卡片......')
@@ -1088,6 +1184,7 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
                                             'size': 'medium',
                                             'icon': {'tag': 'standard_icon', 'token': 'thumbsup_outlined'},
                                             'hover_tips': {'tag': 'plain_text', 'content': '有帮助'},
+                                            'behaviors': [{'type': 'callback', 'value': {'feedback': '有帮助'}}],
                                             'margin': '0px 0px 0px 0px',
                                         }
                                     ],
@@ -1111,6 +1208,7 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
                                             'size': 'medium',
                                             'icon': {'tag': 'standard_icon', 'token': 'thumbdown_outlined'},
                                             'hover_tips': {'tag': 'plain_text', 'content': '无帮助'},
+                                            'behaviors': [{'type': 'callback', 'value': {'feedback': '无帮助'}}],
                                             'margin': '0px 0px 0px 0px',
                                         }
                                     ],
@@ -1190,6 +1288,18 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
             raise Exception(
                 f'client.im.v1.message.reply failed, code: {response.code}, msg: {response.msg}, log_id: {response.get_log_id()}, resp: \n{json.dumps(json.loads(response.raw.content), indent=4, ensure_ascii=False)}'
             )
+
+        # Transfer monitoring message mapping: user msg ID → reply msg ID
+        try:
+            user_msg_id = event.message_chain.message_id
+            reply_msg_id = getattr(response.data, 'message_id', None)
+            monitoring_msg_id = self.pending_monitoring_msg.pop(user_msg_id, None)
+            if reply_msg_id and monitoring_msg_id:
+                self.reply_to_monitoring_msg[reply_msg_id] = (monitoring_msg_id, time.time())
+                self._cleanup_monitoring_mapping()
+        except Exception as e:
+            asyncio.create_task(self.logger.debug(f'Failed to transfer monitoring mapping in create_message_card: {e}'))
+
         return True
 
     async def reply_message(
@@ -1472,6 +1582,58 @@ class LarkAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
 
                 if event.__class__ in self.listeners:
                     await self.listeners[event.__class__](event, self)
+            elif 'card.action.trigger' == type:
+                try:
+                    event_data = data.get('event', {})
+                    operator = event_data.get('operator', {})
+                    action = event_data.get('action', {})
+                    context_data = event_data.get('context', {})
+
+                    action_value_obj = action.get('value', {})
+                    action_value = action_value_obj.get('feedback', '') if isinstance(action_value_obj, dict) else ''
+
+                    if action_value == '有帮助':
+                        feedback_type = 1
+                    elif action_value == '无帮助':
+                        feedback_type = 2
+                    else:
+                        return {'toast': {'type': 'success', 'content': '操作成功'}}
+
+                    user_id = operator.get('open_id') or operator.get('user_id')
+                    open_chat_id = context_data.get('open_chat_id')
+                    open_message_id = context_data.get('open_message_id')
+
+                    if open_chat_id:
+                        session_id = f'group_{open_chat_id}'
+                    elif user_id:
+                        session_id = f'person_{user_id}'
+                    else:
+                        session_id = None
+
+                    # Resolve monitoring message ID from reply message mapping
+                    monitoring_msg_id = None
+                    if open_message_id and open_message_id in self.reply_to_monitoring_msg:
+                        monitoring_msg_id = self.reply_to_monitoring_msg[open_message_id][0]
+
+                    feedback_event = platform_events.FeedbackEvent(
+                        feedback_id=data.get('header', {}).get('event_id', str(uuid.uuid4())),
+                        feedback_type=feedback_type,
+                        feedback_content=action_value,
+                        user_id=user_id,
+                        session_id=session_id,
+                        message_id=open_message_id,
+                        stream_id=monitoring_msg_id,
+                        source_platform_object=data,
+                    )
+
+                    if platform_events.FeedbackEvent in self.listeners:
+                        await self.listeners[platform_events.FeedbackEvent](feedback_event, self)
+
+                    return {'toast': {'type': 'success', 'content': '感谢您的反馈'}}
+                except Exception:
+                    await self.logger.error(f'Error in lark card action callback: {traceback.format_exc()}')
+                    return {'toast': {'type': 'error', 'content': '反馈处理失败'}}
+
             elif 'im.chat.member.bot.added_v1' == type:
                 try:
                     bot_added_welcome_msg = self.config.get('bot_added_welcome', '')

src/langbot/pkg/platform/sources/wecombot.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 import typing
 import asyncio
+import time
 import traceback
 
 import datetime
@@ -126,6 +127,107 @@ class WecomBotMessageConverter(abstract_platform_adapter.AbstractMessageConverte
             if summary:
                 yiri_msg_list.append(platform_message.Plain(text=summary))
 
+        # Handle quoted message (引用消息) - important for group chat file references
+        # Extract files/images/voice from quote and add them as top-level components
+        # so that plugins like FileReader can process them the same way as direct messages
+        quote_info = event.quote or {}
+        if quote_info:
+            # Process quote text content - add as Plain for context
+            if quote_info.get('content'):
+                yiri_msg_list.append(platform_message.Plain(text=f'[引用消息] {quote_info.get("content")}'))
+
+            # Process quote images - add as top-level Image components
+            quote_images = quote_info.get('images', [])
+            if not quote_images and quote_info.get('picurl'):
+                quote_images = [quote_info.get('picurl')]
+            for img_data in quote_images:
+                if img_data:
+                    yiri_msg_list.append(platform_message.Image(base64=img_data))
+
+            # Process quote file - add as top-level File component (same as private chat)
+            quote_file = quote_info.get('file') or {}
+            if quote_file:
+                file_url = (
+                    quote_file.get('base64')
+                    or quote_file.get('download_url')
+                    or quote_file.get('url')
+                    or quote_file.get('fileurl')
+                )
+                file_name = quote_file.get('filename') or quote_file.get('name')
+                file_size = quote_file.get('filesize') or quote_file.get('size')
+                if file_url or file_name:
+                    file_kwargs = {}
+                    if file_url:
+                        file_kwargs['url'] = file_url
+                    if file_name:
+                        file_kwargs['name'] = file_name
+                    if file_size is not None:
+                        file_kwargs['size'] = file_size
+                    try:
+                        yiri_msg_list.append(platform_message.File(**file_kwargs))
+                    except Exception:
+                        yiri_msg_list.append(platform_message.Unknown(text='[quoted file unsupported]'))
+
+            # Process quote voice - add as top-level Voice/File component
+            quote_voice = quote_info.get('voice') or {}
+            if quote_voice:
+                voice_payload = quote_voice.get('base64') or quote_voice.get('url')
+                if voice_payload:
+                    if quote_voice.get('base64') and not voice_payload.startswith('data:'):
+                        voice_payload = f'data:audio/mpeg;base64,{quote_voice.get("base64")}'
+                    try:
+                        yiri_msg_list.append(platform_message.Voice(base64=voice_payload))
+                    except Exception:
+                        try:
+                            voice_kwargs = {'url': voice_payload}
+                            voice_name = quote_voice.get('filename') or quote_voice.get('name')
+                            voice_size = quote_voice.get('filesize') or quote_voice.get('size')
+                            if voice_name:
+                                voice_kwargs['name'] = voice_name
+                            if voice_size is not None:
+                                voice_kwargs['size'] = voice_size
+                            yiri_msg_list.append(platform_message.File(**voice_kwargs))
+                        except Exception:
+                            yiri_msg_list.append(platform_message.Unknown(text='[quoted voice unsupported]'))
+
+            # Process quote video - add as top-level File component
+            quote_video = quote_info.get('video') or {}
+            if quote_video:
+                video_payload = (
+                    quote_video.get('base64')
+                    or quote_video.get('url')
+                    or quote_video.get('download_url')
+                    or quote_video.get('fileurl')
+                )
+                if video_payload:
+                    video_kwargs = {'url': video_payload}
+                    video_name = quote_video.get('filename') or quote_video.get('name')
+                    video_size = quote_video.get('filesize') or quote_video.get('size')
+                    if video_name:
+                        video_kwargs['name'] = video_name
+                    if video_size is not None:
+                        video_kwargs['size'] = video_size
+                    try:
+                        yiri_msg_list.append(platform_message.File(**video_kwargs))
+                    except Exception:
+                        yiri_msg_list.append(platform_message.Unknown(text='[quoted video unsupported]'))
+
+            # Process quote link - add as Plain text
+            quote_link = quote_info.get('link') or {}
+            if quote_link:
+                link_summary = '\n'.join(
+                    filter(
+                        None,
+                        [
+                            quote_link.get('title', ''),
+                            quote_link.get('description') or quote_link.get('digest', ''),
+                            quote_link.get('url', ''),
+                        ],
+                    )
+                )
+                if link_summary:
+                    yiri_msg_list.append(platform_message.Plain(text=f'[引用链接] {link_summary}'))
+
         has_content_element = any(
             not isinstance(element, (platform_message.Source, platform_message.At)) for element in yiri_msg_list
         )
@@ -192,6 +294,8 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
     _ws_mode: bool = False
     bot_name: str = ''
     listeners: dict = {}
+    _stream_to_monitoring_msg: dict = {}  # Maps stream_id to (monitoring_message_id, timestamp)
+    _STREAM_MAPPING_TTL = 600  # 10 minutes
 
     def __init__(self, config: dict, logger: EventLogger):
         enable_webhook = config.get('enable-webhook', False)
@@ -228,8 +332,9 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
             bot_account_id=bot_account_id,
             bot_name=bot_name,
             event_converter=event_converter,
+            listeners={},
+            _stream_to_monitoring_msg={},
         )
-        self.listeners = {}
 
     async def reply_message(
         self,
@@ -321,6 +426,23 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
         """设置 bot UUID（用于生成 webhook URL）"""
         self.bot_uuid = bot_uuid
 
+    async def on_monitoring_message_created(self, query, monitoring_message_id: str):
+        """Called by pipeline after monitoring message is created, to map stream_id to monitoring message ID."""
+        try:
+            stream_id = query.message_event.source_platform_object.stream_id
+            if stream_id:
+                self._stream_to_monitoring_msg[stream_id] = (monitoring_message_id, time.time())
+                self._cleanup_stream_mapping()
+        except Exception as e:
+            await self.logger.debug(f'Failed to map stream_id to monitoring message: {e}')
+
+    def _cleanup_stream_mapping(self):
+        """Remove entries older than TTL from the stream_id to monitoring message mapping."""
+        now = time.time()
+        expired = [k for k, (_, ts) in self._stream_to_monitoring_msg.items() if now - ts > self._STREAM_MAPPING_TTL]
+        for k in expired:
+            del self._stream_to_monitoring_msg[k]
+
     async def _on_feedback(self, **kwargs):
         """Handle feedback event from WeChat Work AI Bot SDK and dispatch as FeedbackEvent."""
         try:
@@ -328,6 +450,9 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
             feedback_type = kwargs.get('feedback_type', 0)
             feedback_content = kwargs.get('feedback_content', '') or None
             inaccurate_reasons = kwargs.get('inaccurate_reasons', []) or None
+            # WeChat Work returns integer reason codes, but FeedbackEvent expects strings
+            if inaccurate_reasons:
+                inaccurate_reasons = [str(r) for r in inaccurate_reasons]
             session = kwargs.get('session')
 
             session_id = None
@@ -343,6 +468,16 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
                 message_id = session.msg_id
                 stream_id = session.stream_id
 
+            # Resolve stream_id to LangBot monitoring message ID if available
+            monitoring_msg_id = None
+            if stream_id and stream_id in self._stream_to_monitoring_msg:
+                monitoring_msg_id = self._stream_to_monitoring_msg[stream_id][0]
+
+            await self.logger.info(
+                f'Feedback event: feedback_id={feedback_id}, type={feedback_type}, '
+                f'session_id={session_id}, user_id={user_id}, message_id={message_id}'
+            )
+
             event = platform_events.FeedbackEvent(
                 feedback_id=feedback_id,
                 feedback_type=feedback_type,
@@ -351,7 +486,7 @@ class WecomBotAdapter(abstract_platform_adapter.AbstractMessagePlatformAdapter):
                 user_id=user_id,
                 session_id=session_id,
                 message_id=message_id,
-                stream_id=stream_id,
+                stream_id=monitoring_msg_id or stream_id,
                 source_platform_object=session,
             )
 

src/langbot/pkg/provider/modelmgr/modelmgr.py

@@ -9,7 +9,6 @@ from ...discover import engine
 from . import token
 from ...entity.persistence import model as persistence_model
 from ...entity.errors import provider as provider_errors
-from async_lru import alru_cache
 
 
 class ModelManager:
@@ -24,6 +23,8 @@ class ModelManager:
 
     embedding_models: list[requester.RuntimeEmbeddingModel]
 
+    rerank_models: list[requester.RuntimeRerankModel]
+
     requester_components: list[engine.Component]
 
     requester_dict: dict[str, type[requester.ProviderAPIRequester]]
@@ -32,6 +33,7 @@ class ModelManager:
         self.ap = ap
         self.llm_models = []
         self.embedding_models = []
+        self.rerank_models = []
         self.requester_components = []
         self.requester_dict = {}
 
@@ -64,8 +66,7 @@ class ModelManager:
 
         self.llm_models = []
         self.embedding_models = []
-
-        # Load all providers first
+        self.rerank_models = []
         self.provider_dict = {}
         providers_result = await self.ap.persistence_mgr.execute_async(
             sqlalchemy.select(persistence_model.ModelProvider)
@@ -110,6 +111,22 @@ class ModelManager:
             except Exception as e:
                 self.ap.logger.error(f'Failed to load model {embedding_model.uuid}: {e}\n{traceback.format_exc()}')
 
+        # Load rerank models
+        result = await self.ap.persistence_mgr.execute_async(sqlalchemy.select(persistence_model.RerankModel))
+        rerank_models = result.all()
+        for rerank_model in rerank_models:
+            try:
+                provider = self.provider_dict.get(rerank_model.provider_uuid)
+                if provider is None:
+                    self.ap.logger.warning(
+                        f'Provider {rerank_model.provider_uuid} not found for model {rerank_model.uuid}'
+                    )
+                    continue
+                runtime_rerank_model = await self.load_rerank_model_with_provider(rerank_model, provider)
+                self.rerank_models.append(runtime_rerank_model)
+            except Exception as e:
+                self.ap.logger.error(f'Failed to load model {rerank_model.uuid}: {e}\n{traceback.format_exc()}')
+
     async def sync_new_models_from_space(self):
         """Sync models from Space"""
         space_model_provider = await self.ap.persistence_mgr.execute_async(
@@ -212,6 +229,26 @@ class ModelManager:
 
         return runtime_embedding_model
 
+    async def init_temporary_runtime_rerank_model(
+        self,
+        model_info: dict,
+    ) -> requester.RuntimeRerankModel:
+        """Initialize runtime rerank model from dict (for testing)"""
+        provider_info = model_info.get('provider', {})
+        runtime_provider = await self.load_provider(provider_info)
+
+        runtime_rerank_model = requester.RuntimeRerankModel(
+            model_entity=persistence_model.RerankModel(
+                uuid=model_info.get('uuid', ''),
+                name=model_info.get('name', ''),
+                provider_uuid='',
+                extra_args=model_info.get('extra_args', {}),
+            ),
+            provider=runtime_provider,
+        )
+
+        return runtime_rerank_model
+
     async def load_provider(
         self, provider_info: persistence_model.ModelProvider | sqlalchemy.Row | dict
     ) -> requester.RuntimeProvider:
@@ -227,7 +264,8 @@ class ModelManager:
             raise provider_errors.RequesterNotFoundError(provider_entity.requester)
 
         requester_inst = self.requester_dict[provider_entity.requester](
-            ap=self.ap, config={'base_url': provider_entity.base_url}
+            ap=self.ap,
+            config={'base_url': provider_entity.base_url},
         )
         await requester_inst.initialize()
 
@@ -268,6 +306,9 @@ class ModelManager:
         for model in self.embedding_models:
             if model.provider.provider_entity.uuid == provider_uuid:
                 model.provider = new_runtime_provider
+        for model in self.rerank_models:
+            if model.provider.provider_entity.uuid == provider_uuid:
+                model.provider = new_runtime_provider
 
         # update ref in provider dict
         self.provider_dict[provider_uuid] = new_runtime_provider
@@ -304,6 +345,22 @@ class ModelManager:
 
         return runtime_embedding_model
 
+    async def load_rerank_model_with_provider(
+        self,
+        model_info: persistence_model.RerankModel | sqlalchemy.Row,
+        provider: requester.RuntimeProvider,
+    ) -> requester.RuntimeRerankModel:
+        """Load rerank model with provider info"""
+        if isinstance(model_info, sqlalchemy.Row):
+            model_info = persistence_model.RerankModel(**model_info._mapping)
+
+        runtime_rerank_model = requester.RuntimeRerankModel(
+            model_entity=model_info,
+            provider=provider,
+        )
+
+        return runtime_rerank_model
+
     async def load_llm_model(self, model_info: dict):
         """Load LLM model from dict (with provider info)"""
         provider_info = model_info.get('provider', {})
@@ -351,7 +408,6 @@ class ModelManager:
 
         await self.load_embedding_model_with_provider(model_entity, provider_entity)
 
-    @alru_cache(ttl=60 * 5)
     async def get_model_by_uuid(self, uuid: str) -> requester.RuntimeLLMModel:
         """Get LLM model by uuid"""
         for model in self.llm_models:
@@ -359,7 +415,6 @@ class ModelManager:
                 return model
         raise ValueError(f'LLM model {uuid} not found')
 
-    @alru_cache(ttl=60 * 5)
     async def get_embedding_model_by_uuid(self, uuid: str) -> requester.RuntimeEmbeddingModel:
         """Get embedding model by uuid"""
         for model in self.embedding_models:
@@ -367,6 +422,13 @@ class ModelManager:
                 return model
         raise ValueError(f'Embedding model {uuid} not found')
 
+    async def get_rerank_model_by_uuid(self, uuid: str) -> requester.RuntimeRerankModel:
+        """Get rerank model by uuid"""
+        for model in self.rerank_models:
+            if model.model_entity.uuid == uuid:
+                return model
+        raise ValueError(f'Rerank model {uuid} not found')
+
     async def remove_llm_model(self, model_uuid: str):
         """Remove LLM model"""
         for model in self.llm_models:
@@ -381,6 +443,13 @@ class ModelManager:
                 self.embedding_models.remove(model)
                 return
 
+    async def remove_rerank_model(self, model_uuid: str):
+        """Remove rerank model"""
+        for model in self.rerank_models:
+            if model.model_entity.uuid == model_uuid:
+                self.rerank_models.remove(model)
+                return
+
     def get_available_requesters_info(self, model_type: str) -> list[dict]:
         """Get all available requesters"""
         if model_type != '':

src/langbot/pkg/provider/modelmgr/requester.py

@@ -247,6 +247,40 @@ class RuntimeProvider:
             except Exception as monitor_err:
                 self.requester.ap.logger.error(f'[Monitoring] Failed to record embedding call: {monitor_err}')
 
+    async def invoke_rerank(
+        self,
+        model: RuntimeRerankModel,
+        query: str,
+        documents: typing.List[str],
+        extra_args: dict[str, typing.Any] = {},
+    ) -> typing.List[dict]:
+        """Bridge method for invoking rerank with monitoring"""
+        start_time = time.time()
+        status = 'success'
+
+        try:
+            result = await self.requester.invoke_rerank(
+                model=model,
+                query=query,
+                documents=documents,
+                extra_args=extra_args,
+            )
+            return result
+
+        except Exception:
+            status = 'error'
+            raise
+        finally:
+            duration_ms = int((time.time() - start_time) * 1000)
+
+            try:
+                self.requester.ap.logger.debug(
+                    f'[Rerank] model={model.model_entity.name} docs={len(documents)} '
+                    f'duration={duration_ms}ms status={status}'
+                )
+            except Exception as monitor_err:
+                self.requester.ap.logger.error(f'[Monitoring] Failed to record rerank call: {monitor_err}')
+
 
 class RuntimeLLMModel:
     """运行时模型"""
@@ -284,6 +318,24 @@ class RuntimeEmbeddingModel:
         self.provider = provider
 
 
+class RuntimeRerankModel:
+    """运行时 Rerank 模型"""
+
+    model_entity: persistence_model.RerankModel
+    """模型数据"""
+
+    provider: RuntimeProvider
+    """提供商实例"""
+
+    def __init__(
+        self,
+        model_entity: persistence_model.RerankModel,
+        provider: RuntimeProvider,
+    ):
+        self.model_entity = model_entity
+        self.provider = provider
+
+
 class ProviderAPIRequester(metaclass=abc.ABCMeta):
     """Provider API请求器"""
 
@@ -303,6 +355,14 @@ class ProviderAPIRequester(metaclass=abc.ABCMeta):
     async def initialize(self):
         pass
 
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any] | list[dict[str, typing.Any]]:
+        """Scan models supported by the provider.
+
+        The default implementation does not support scanning. Requesters that
+        can enumerate remote models should override this method.
+        """
+        raise NotImplementedError('This provider does not support model scanning')
+
     @abc.abstractmethod
     async def invoke_llm(
         self,
@@ -368,3 +428,23 @@ class ProviderAPIRequester(metaclass=abc.ABCMeta):
             或者 tuple[typing.List[typing.List[float]], dict]: 返回 (embedding 向量, usage_info)
         """
         pass
+
+    async def invoke_rerank(
+        self,
+        model: RuntimeRerankModel,
+        query: str,
+        documents: typing.List[str],
+        extra_args: dict[str, typing.Any] = {},
+    ) -> typing.List[dict]:
+        """调用 Rerank API
+
+        Args:
+            model (RuntimeRerankModel): 使用的模型信息
+            query (str): 查询文本
+            documents (typing.List[str]): 待重排序的文档列表
+            extra_args (dict[str, typing.Any], optional): 额外的参数. Defaults to {}.
+
+        Returns:
+            typing.List[dict]: [{"index": int, "relevance_score": float}, ...]
+        """
+        raise NotImplementedError('This requester does not support rerank')

src/langbot/pkg/provider/modelmgr/requesters/302aichatcmpl.yaml

@@ -25,6 +25,7 @@ spec:
   support_type:
   - llm
   - text-embedding
+  - rerank
   provider_category: maas
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/bailianchatcmpl.yaml

@@ -24,6 +24,7 @@ spec:
     default: 120
   support_type:
   - llm
+  - rerank
   provider_category: maas
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/chatcmpl.py

@@ -31,6 +31,192 @@ class OpenAIChatCompletions(requester.ProviderAPIRequester):
             http_client=httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']),
         )
 
+    def _mask_api_key(self, api_key: str | None) -> str:
+        if not api_key:
+            return ''
+        if len(api_key) <= 8:
+            return '****'
+        return f'{api_key[:4]}...{api_key[-4:]}'
+
+    def _infer_model_type(self, model_id: str) -> str:
+        normalized_model_id = (model_id or '').lower()
+        embedding_keywords = (
+            'embedding',
+            'embed',
+            'bge-',
+            'e5-',
+            'm3e',
+            'gte-',
+            'multilingual-e5',
+            'text-embedding',
+        )
+        return 'embedding' if any(keyword in normalized_model_id for keyword in embedding_keywords) else 'llm'
+
+    def _infer_model_abilities(self, item: dict[str, typing.Any], model_id: str) -> list[str]:
+        normalized_model_id = (model_id or '').lower()
+        abilities: set[str] = set()
+
+        def _flatten(value: typing.Any) -> list[str]:
+            if value is None:
+                return []
+            if isinstance(value, str):
+                return [value.lower()]
+            if isinstance(value, dict):
+                flattened: list[str] = []
+                for nested_value in value.values():
+                    flattened.extend(_flatten(nested_value))
+                return flattened
+            if isinstance(value, (list, tuple, set)):
+                flattened: list[str] = []
+                for nested_value in value:
+                    flattened.extend(_flatten(nested_value))
+                return flattened
+            return [str(value).lower()]
+
+        capability_tokens = _flatten(item.get('capabilities'))
+        capability_tokens.extend(_flatten(item.get('modalities')))
+        capability_tokens.extend(_flatten(item.get('input_modalities')))
+        capability_tokens.extend(_flatten(item.get('output_modalities')))
+        capability_tokens.extend(_flatten(item.get('supported_generation_methods')))
+        capability_tokens.extend(_flatten(item.get('supported_parameters')))
+        capability_tokens.extend(_flatten(item.get('architecture')))
+
+        combined_tokens = capability_tokens + [normalized_model_id]
+
+        vision_keywords = (
+            'vision',
+            'image',
+            'file',
+            'video',
+            'multimodal',
+            'vl',
+            'ocr',
+            'omni',
+        )
+        function_call_keywords = (
+            'function',
+            'tool',
+            'tools',
+            'tool_choice',
+            'tool_call',
+            'tool-use',
+            'tool_use',
+        )
+
+        if any(any(keyword in token for keyword in vision_keywords) for token in combined_tokens):
+            abilities.add('vision')
+
+        if any(any(keyword in token for keyword in function_call_keywords) for token in combined_tokens):
+            abilities.add('func_call')
+
+        return sorted(abilities)
+
+    def _normalize_modalities(self, value: typing.Any) -> list[str]:
+        normalized: list[str] = []
+
+        def _collect(item: typing.Any):
+            if item is None:
+                return
+            if isinstance(item, str):
+                for part in item.replace('->', ',').replace('+', ',').split(','):
+                    token = part.strip().lower()
+                    if token and token not in normalized:
+                        normalized.append(token)
+                return
+            if isinstance(item, dict):
+                for nested in item.values():
+                    _collect(nested)
+                return
+            if isinstance(item, (list, tuple, set)):
+                for nested in item:
+                    _collect(nested)
+                return
+
+        _collect(value)
+        return normalized
+
+    def _extract_scan_metadata(self, item: dict[str, typing.Any], model_id: str) -> dict[str, typing.Any]:
+        display_name = item.get('name')
+        if not isinstance(display_name, str) or not display_name.strip() or display_name == model_id:
+            display_name = ''
+
+        description = item.get('description')
+        if not isinstance(description, str) or not description.strip():
+            description = ''
+
+        context_length = item.get('context_length')
+        if context_length is None and isinstance(item.get('top_provider'), dict):
+            context_length = item['top_provider'].get('context_length')
+
+        if not isinstance(context_length, int):
+            try:
+                context_length = int(context_length) if context_length is not None else None
+            except (TypeError, ValueError):
+                context_length = None
+
+        input_modalities = self._normalize_modalities(item.get('input_modalities'))
+        output_modalities = self._normalize_modalities(item.get('output_modalities'))
+
+        if isinstance(item.get('architecture'), dict):
+            if not input_modalities:
+                input_modalities = self._normalize_modalities(item['architecture'].get('input_modalities'))
+            if not output_modalities:
+                output_modalities = self._normalize_modalities(item['architecture'].get('output_modalities'))
+
+        owned_by = item.get('owned_by')
+        if not isinstance(owned_by, str) or not owned_by.strip():
+            owned_by = ''
+
+        return {
+            'display_name': display_name or None,
+            'description': description or None,
+            'context_length': context_length,
+            'owned_by': owned_by or None,
+            'input_modalities': input_modalities,
+            'output_modalities': output_modalities,
+        }
+
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any]:
+        headers = {}
+        if api_key:
+            headers['Authorization'] = f'Bearer {api_key}'
+
+        models_url = f'{self.requester_cfg["base_url"].rstrip("/")}/models'
+        async with httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']) as client:
+            response = await client.get(models_url, headers=headers)
+            response.raise_for_status()
+            payload = response.json()
+
+        models = []
+        for item in payload.get('data', []):
+            model_id = item.get('id')
+            if not model_id:
+                continue
+            models.append(
+                {
+                    'id': model_id,
+                    'name': model_id,
+                    'type': self._infer_model_type(model_id),
+                    'abilities': self._infer_model_abilities(item, model_id),
+                    **self._extract_scan_metadata(item, model_id),
+                }
+            )
+
+        models.sort(key=lambda item: (item['type'] != 'llm', item['name'].lower()))
+        return {
+            'models': models,
+            'debug': {
+                'request': {
+                    'method': 'GET',
+                    'url': models_url,
+                    'headers': {
+                        'Authorization': f'Bearer {self._mask_api_key(api_key)}' if api_key else '',
+                    },
+                },
+                'response': payload,
+            },
+        }
+
     async def _req(
         self,
         args: dict,
@@ -429,3 +615,88 @@ class OpenAIChatCompletions(requester.ProviderAPIRequester):
             raise errors.RequesterError(f'请求过于频繁或余额不足: {e.message}')
         except openai.APIError as e:
             raise errors.RequesterError(f'请求错误: {e.message}')
+
+    async def invoke_rerank(
+        self,
+        model: requester.RuntimeRerankModel,
+        query: str,
+        documents: typing.List[str],
+        extra_args: dict[str, typing.Any] = {},
+    ) -> typing.List[dict]:
+        """Standard /rerank endpoint (Jina/Cohere/SiliconFlow/Voyage/DashScope compatible)
+
+        Supports extra_args from model.extra_args:
+        - rerank_url: full URL override (e.g. "https://dashscope.aliyuncs.com/compatible-api/v1/reranks")
+        - rerank_path: path override appended to base_url (e.g. "reranks" instead of default "rerank")
+        - Any other fields are merged into the request payload.
+        """
+        api_key = model.provider.token_mgr.get_token()
+        base_url = self.requester_cfg.get('base_url', '').rstrip('/')
+        timeout = self.requester_cfg.get('timeout', 120)
+
+        merged_args = {}
+        if model.model_entity.extra_args:
+            merged_args.update(model.model_entity.extra_args)
+        if extra_args:
+            merged_args.update(extra_args)
+
+        rerank_url = merged_args.pop('rerank_url', None)
+        rerank_path = merged_args.pop('rerank_path', 'rerank')
+        if not rerank_url:
+            rerank_url = f'{base_url}/{rerank_path}'
+
+        headers = {
+            'Content-Type': 'application/json',
+            'Authorization': f'Bearer {api_key}',
+        }
+
+        payload = {
+            'model': model.model_entity.name,
+            'query': query,
+            'documents': documents[:64],
+            'top_n': min(len(documents), 64),
+        }
+
+        if merged_args:
+            payload.update(merged_args)
+
+        try:
+            async with httpx.AsyncClient(trust_env=True, timeout=timeout) as client:
+                resp = await client.post(rerank_url, headers=headers, json=payload)
+                resp.raise_for_status()
+                data = resp.json()
+
+            results = self._parse_rerank_response(data)
+
+            if results:
+                scores = [r.get('relevance_score', 0.0) for r in results]
+                min_score = min(scores)
+                max_score = max(scores)
+                if max_score - min_score > 1e-6:
+                    for r in results:
+                        r['relevance_score'] = (r['relevance_score'] - min_score) / (max_score - min_score)
+
+            return results
+        except httpx.HTTPStatusError as e:
+            raise errors.RequesterError(f'Rerank request failed: {e.response.status_code} - {e.response.text}')
+        except httpx.TimeoutException:
+            raise errors.RequesterError('Rerank request timed out')
+        except Exception as e:
+            raise errors.RequesterError(f'Rerank request error: {str(e)}')
+
+    @staticmethod
+    def _parse_rerank_response(data: dict) -> typing.List[dict]:
+        """Parse rerank response from various providers.
+
+        Handles:
+        - Jina/Cohere/SiliconFlow: {"results": [{"index", "relevance_score"}]}
+        - Voyage AI: {"data": [{"index", "relevance_score"}]}
+        - DashScope: {"output": {"results": [{"index", "relevance_score"}]}}
+        """
+        if 'results' in data:
+            return data['results']
+        if 'data' in data:
+            return data['data']
+        if 'output' in data and isinstance(data['output'], dict):
+            return data['output'].get('results', [])
+        return []

src/langbot/pkg/provider/modelmgr/requesters/chatcmpl.yaml

@@ -25,6 +25,7 @@ spec:
   support_type:
   - llm
   - text-embedding
+  - rerank
   provider_category: manufacturer
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/chroma.svg

@@ -0,0 +1,8 @@
+<svg xmlns="http://www.w3.org/2000/svg" fill="none" viewBox="0 0 128 128" id="Chroma--Streamline-Svg-Logos" height="128" width="128">
+  <desc>
+    Chroma Streamline Icon: https://streamlinehq.com
+  </desc>
+  <path fill="#ffde2d" d="M84.88839999999999 104.10666666666665c23.0732 0 41.77773333333333 -17.956266666666664 41.77773333333333 -40.10653333333333 0 -22.150266666666667 -18.70453333333333 -40.10653333333333 -41.77773333333333 -40.10653333333333 -23.0732 0 -41.77773333333333 17.956266666666664 -41.77773333333333 40.10653333333333 0 22.150266666666667 18.70453333333333 40.10653333333333 41.77773333333333 40.10653333333333Z" stroke-width="1.3333"></path>
+  <path fill="#327eff" d="M43.111066666666666 104.10666666666665c23.0732 0 41.77773333333333 -17.956266666666664 41.77773333333333 -40.10653333333333 0 -22.150266666666667 -18.70453333333333 -40.10653333333333 -41.77773333333333 -40.10653333333333C20.037866666666666 23.8936 1.3333333333333333 41.849866666666664 1.3333333333333333 64.00013333333334 1.3333333333333333 86.15039999999999 20.037866666666666 104.10666666666665 43.111066666666666 104.10666666666665Z" stroke-width="1.3333"></path>
+  <path fill="#ff6446" d="M84.88866666666667 64.00013333333334c0 22.150399999999998 -18.704666666666665 40.10626666666666 -41.778 40.10626666666666V64.00013333333334h41.778Zm-41.778 0c0 -22.150266666666667 18.70453333333333 -40.10653333333333 41.778 -40.10653333333333v40.10653333333333H43.11066666666666Z" stroke-width="1.3333"></path>
+</svg>

src/langbot/pkg/provider/modelmgr/requesters/chromaembed.py

@@ -0,0 +1,61 @@
+from __future__ import annotations
+
+import typing
+
+from .. import requester
+
+REQUESTER_NAME: str = 'chroma-embedding'
+
+
+class ChromaEmbedding(requester.ProviderAPIRequester):
+    """Chroma built-in embedding requester.
+
+    Uses chromadb's DefaultEmbeddingFunction (all-MiniLM-L6-v2).
+    The embedding function runs locally using ONNX Runtime.
+    """
+
+    default_config: dict[str, typing.Any] = {
+        'base_url': '',
+    }
+
+    _embedding_function = None
+
+    async def initialize(self):
+        try:
+            from chromadb.utils import embedding_functions
+        except ImportError:
+            raise ImportError('chromadb is not installed. Install it with: pip install chromadb')
+
+        self._embedding_function = embedding_functions.DefaultEmbeddingFunction()
+
+    async def invoke_llm(
+        self,
+        query,
+        model: requester.RuntimeLLMModel,
+        messages: typing.List,
+        funcs: typing.List = None,
+        extra_args: dict[str, typing.Any] = {},
+        remove_think: bool = False,
+    ):
+        raise NotImplementedError('Chroma embedding does not support LLM inference')
+
+    async def invoke_embedding(
+        self,
+        model: requester.RuntimeEmbeddingModel,
+        input_text: typing.List[str],
+        extra_args: dict[str, typing.Any] = {},
+    ) -> typing.List[typing.List[float]]:
+        """Generate embeddings using Chroma's DefaultEmbeddingFunction."""
+        if self._embedding_function is None:
+            await self.initialize()
+
+        try:
+            result = self._embedding_function(input_text)
+            # DefaultEmbeddingFunction returns list of ndarray, convert for JSON
+            if isinstance(result, list):
+                return [item.tolist() if hasattr(item, 'tolist') else item for item in result]
+            return result.tolist() if hasattr(result, 'tolist') else result
+        except Exception as e:
+            from .. import errors
+
+            raise errors.RequesterError(f'Chroma embedding failed: {str(e)}')

src/langbot/pkg/provider/modelmgr/requesters/chromaembed.yaml

@@ -0,0 +1,21 @@
+apiVersion: v1
+kind: LLMAPIRequester
+metadata:
+  name: chroma-embedding
+  label:
+    en_US: Chroma Embedding
+    zh_Hans: Chroma 嵌入
+  description:
+    en_US: Chroma built-in embedding model (all-MiniLM-L6-v2), runs locally using ONNX Runtime. First-time use will download model files automatically.
+    zh_Hans: 使用 Chroma 内置嵌入模型 (all-MiniLM-L6-v2)，基于 ONNX Runtime 本地运行。首次使用时将自动下载模型文件。
+    ja_JP: Chroma 組み込み埋め込みモデル (all-MiniLM-L6-v2) を使用します。ONNX Runtime でローカル実行。初回使用時にモデルファイルが自動ダウンロードされます。
+  icon: chroma.svg
+spec:
+  config: []
+  support_type:
+  - text-embedding
+  provider_category: builtin
+execution:
+  python:
+    path: ./chromaembed.py
+    attr: ChromaEmbedding

src/langbot/pkg/provider/modelmgr/requesters/cohere.svg

@@ -0,0 +1 @@
+<svg height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>Cohere</title><path clip-rule="evenodd" d="M8.128 14.099c.592 0 1.77-.033 3.398-.703 1.897-.781 5.672-2.2 8.395-3.656 1.905-1.018 2.74-2.366 2.74-4.18A4.56 4.56 0 0018.1 1H7.549A6.55 6.55 0 001 7.55c0 3.617 2.745 6.549 7.128 6.549z" fill="#39594D" fill-rule="evenodd"></path><path clip-rule="evenodd" d="M9.912 18.61a4.387 4.387 0 012.705-4.052l3.323-1.38c3.361-1.394 7.06 1.076 7.06 4.715a5.104 5.104 0 01-5.105 5.104l-3.597-.001a4.386 4.386 0 01-4.386-4.387z" fill="#D18EE2" fill-rule="evenodd"></path><path d="M4.776 14.962A3.775 3.775 0 001 18.738v.489a3.776 3.776 0 007.551 0v-.49a3.775 3.775 0 00-3.775-3.775z" fill="#FF7759"></path></svg>

src/langbot/pkg/provider/modelmgr/requesters/coherererank.yaml

@@ -0,0 +1,31 @@
+apiVersion: v1
+kind: LLMAPIRequester
+metadata:
+  name: cohere-rerank
+  label:
+    en_US: Cohere
+    zh_Hans: Cohere
+  icon: cohere.svg
+spec:
+  config:
+  - name: base_url
+    label:
+      en_US: Base URL
+      zh_Hans: 基础 URL
+    type: string
+    required: true
+    default: https://api.cohere.com/v2
+  - name: timeout
+    label:
+      en_US: Timeout
+      zh_Hans: 超时时间
+    type: integer
+    required: true
+    default: 120
+  support_type:
+  - rerank
+  provider_category: manufacturer
+execution:
+  python:
+    path: ./chatcmpl.py
+    attr: OpenAIChatCompletions

src/langbot/pkg/provider/modelmgr/requesters/geminichatcmpl.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import typing
+import httpx
 
 from . import chatcmpl
 
@@ -20,6 +21,68 @@ class GeminiChatCompletions(chatcmpl.OpenAIChatCompletions):
         'timeout': 120,
     }
 
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any]:
+        models_url = 'https://generativelanguage.googleapis.com/v1beta/models'
+        params = {'key': api_key} if api_key else {}
+
+        all_models: list[dict[str, typing.Any]] = []
+        next_page_token = ''
+        last_payload: dict[str, typing.Any] = {}
+
+        async with httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']) as client:
+            while True:
+                request_params = dict(params)
+                if next_page_token:
+                    request_params['pageToken'] = next_page_token
+
+                response = await client.get(models_url, params=request_params)
+                response.raise_for_status()
+                payload = response.json()
+                last_payload = payload
+
+                for item in payload.get('models', []):
+                    model_name = item.get('name', '')
+                    model_id = model_name.replace('models/', '', 1)
+                    if not model_id:
+                        continue
+
+                    supported_methods = item.get('supportedGenerationMethods', []) or []
+                    if 'embedContent' in supported_methods and 'generateContent' not in supported_methods:
+                        model_type = 'embedding'
+                    else:
+                        model_type = 'llm'
+
+                    all_models.append(
+                        {
+                            'id': model_id,
+                            'name': model_id,
+                            'type': model_type,
+                            'abilities': self._infer_model_abilities(item, model_id),
+                            'display_name': item.get('displayName') or None,
+                            'description': item.get('description') or None,
+                            'context_length': item.get('inputTokenLimit'),
+                            'input_modalities': self._normalize_modalities(item.get('inputModalities')),
+                            'output_modalities': self._normalize_modalities(item.get('outputModalities')),
+                        }
+                    )
+
+                next_page_token = payload.get('nextPageToken', '')
+                if not next_page_token:
+                    break
+
+        all_models.sort(key=lambda item: (item['type'] != 'llm', item['name'].lower()))
+        return {
+            'models': all_models,
+            'debug': {
+                'request': {
+                    'method': 'GET',
+                    'url': models_url,
+                    'query': {'key': self._mask_api_key(api_key)} if api_key else {},
+                },
+                'response': last_payload,
+            },
+        }
+
     async def _closure_stream(
         self,
         query: pipeline_query.Query,

src/langbot/pkg/provider/modelmgr/requesters/giteeaichatcmpl.yaml

@@ -25,6 +25,7 @@ spec:
   support_type:
   - llm
   - text-embedding
+  - rerank
   provider_category: maas
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/jina.svg

@@ -0,0 +1 @@
+<svg fill="currentColor" fill-rule="evenodd" height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>Jina</title><path d="M6.608 21.416a4.608 4.608 0 100-9.217 4.608 4.608 0 000 9.217zM20.894 2.015c.614 0 1.106.492 1.106 1.106v9.002c0 5.13-4.148 9.309-9.217 9.37v-9.355l-.03-9.032c0-.614.491-1.106 1.106-1.106h7.158l-.123.015z"></path></svg>

src/langbot/pkg/provider/modelmgr/requesters/jinarerank.yaml

@@ -0,0 +1,31 @@
+apiVersion: v1
+kind: LLMAPIRequester
+metadata:
+  name: jina-rerank
+  label:
+    en_US: Jina
+    zh_Hans: Jina
+  icon: jina.svg
+spec:
+  config:
+  - name: base_url
+    label:
+      en_US: Base URL
+      zh_Hans: 基础 URL
+    type: string
+    required: true
+    default: https://api.jina.ai/v1
+  - name: timeout
+    label:
+      en_US: Timeout
+      zh_Hans: 超时时间
+    type: integer
+    required: true
+    default: 120
+  support_type:
+  - rerank
+  provider_category: manufacturer
+execution:
+  python:
+    path: ./chatcmpl.py
+    attr: OpenAIChatCompletions

src/langbot/pkg/provider/modelmgr/requesters/modelscopechatcmpl.py

@@ -31,6 +31,175 @@ class ModelScopeChatCompletions(requester.ProviderAPIRequester):
             http_client=httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']),
         )
 
+    def _mask_api_key(self, api_key: str | None) -> str:
+        if not api_key:
+            return ''
+        if len(api_key) <= 8:
+            return '****'
+        return f'{api_key[:4]}...{api_key[-4:]}'
+
+    def _infer_model_type(self, model_id: str) -> str:
+        normalized_model_id = (model_id or '').lower()
+        embedding_keywords = (
+            'embedding',
+            'embed',
+            'bge-',
+            'e5-',
+            'm3e',
+            'gte-',
+            'multilingual-e5',
+            'text-embedding',
+        )
+        return 'embedding' if any(keyword in normalized_model_id for keyword in embedding_keywords) else 'llm'
+
+    def _infer_model_abilities(self, item: dict[str, typing.Any], model_id: str) -> list[str]:
+        normalized_model_id = (model_id or '').lower()
+        abilities: set[str] = set()
+
+        def _flatten(value: typing.Any) -> list[str]:
+            if value is None:
+                return []
+            if isinstance(value, str):
+                return [value.lower()]
+            if isinstance(value, dict):
+                flattened: list[str] = []
+                for nested_value in value.values():
+                    flattened.extend(_flatten(nested_value))
+                return flattened
+            if isinstance(value, (list, tuple, set)):
+                flattened: list[str] = []
+                for nested_value in value:
+                    flattened.extend(_flatten(nested_value))
+                return flattened
+            return [str(value).lower()]
+
+        capability_tokens = _flatten(item.get('capabilities'))
+        capability_tokens.extend(_flatten(item.get('modalities')))
+        capability_tokens.extend(_flatten(item.get('input_modalities')))
+        capability_tokens.extend(_flatten(item.get('output_modalities')))
+        capability_tokens.extend(_flatten(item.get('supported_generation_methods')))
+        capability_tokens.extend(_flatten(item.get('supported_parameters')))
+        capability_tokens.extend(_flatten(item.get('architecture')))
+
+        combined_tokens = capability_tokens + [normalized_model_id]
+
+        vision_keywords = ('vision', 'image', 'file', 'video', 'multimodal', 'vl', 'ocr', 'omni')
+        function_call_keywords = ('function', 'tool', 'tools', 'tool_choice', 'tool_call', 'tool-use', 'tool_use')
+
+        if any(any(keyword in token for keyword in vision_keywords) for token in combined_tokens):
+            abilities.add('vision')
+
+        if any(any(keyword in token for keyword in function_call_keywords) for token in combined_tokens):
+            abilities.add('func_call')
+
+        return sorted(abilities)
+
+    def _normalize_modalities(self, value: typing.Any) -> list[str]:
+        normalized: list[str] = []
+
+        def _collect(item: typing.Any):
+            if item is None:
+                return
+            if isinstance(item, str):
+                for part in item.replace('->', ',').replace('+', ',').split(','):
+                    token = part.strip().lower()
+                    if token and token not in normalized:
+                        normalized.append(token)
+                return
+            if isinstance(item, dict):
+                for nested in item.values():
+                    _collect(nested)
+                return
+            if isinstance(item, (list, tuple, set)):
+                for nested in item:
+                    _collect(nested)
+                return
+
+        _collect(value)
+        return normalized
+
+    def _extract_scan_metadata(self, item: dict[str, typing.Any], model_id: str) -> dict[str, typing.Any]:
+        display_name = item.get('name')
+        if not isinstance(display_name, str) or not display_name.strip() or display_name == model_id:
+            display_name = ''
+
+        description = item.get('description')
+        if not isinstance(description, str) or not description.strip():
+            description = ''
+
+        context_length = item.get('context_length')
+        if context_length is None and isinstance(item.get('top_provider'), dict):
+            context_length = item['top_provider'].get('context_length')
+
+        if not isinstance(context_length, int):
+            try:
+                context_length = int(context_length) if context_length is not None else None
+            except (TypeError, ValueError):
+                context_length = None
+
+        input_modalities = self._normalize_modalities(item.get('input_modalities'))
+        output_modalities = self._normalize_modalities(item.get('output_modalities'))
+
+        if isinstance(item.get('architecture'), dict):
+            if not input_modalities:
+                input_modalities = self._normalize_modalities(item['architecture'].get('input_modalities'))
+            if not output_modalities:
+                output_modalities = self._normalize_modalities(item['architecture'].get('output_modalities'))
+
+        owned_by = item.get('owned_by')
+        if not isinstance(owned_by, str) or not owned_by.strip():
+            owned_by = ''
+
+        return {
+            'display_name': display_name or None,
+            'description': description or None,
+            'context_length': context_length,
+            'owned_by': owned_by or None,
+            'input_modalities': input_modalities,
+            'output_modalities': output_modalities,
+        }
+
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any]:
+        headers = {}
+        if api_key:
+            headers['Authorization'] = f'Bearer {api_key}'
+
+        models_url = f'{self.requester_cfg["base_url"].rstrip("/")}/models'
+        async with httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']) as client:
+            response = await client.get(models_url, headers=headers)
+            response.raise_for_status()
+            payload = response.json()
+
+        models = []
+        for item in payload.get('data', []):
+            model_id = item.get('id')
+            if not model_id:
+                continue
+            models.append(
+                {
+                    'id': model_id,
+                    'name': model_id,
+                    'type': self._infer_model_type(model_id),
+                    'abilities': self._infer_model_abilities(item, model_id),
+                    **self._extract_scan_metadata(item, model_id),
+                }
+            )
+
+        models.sort(key=lambda item: (item['type'] != 'llm', item['name'].lower()))
+        return {
+            'models': models,
+            'debug': {
+                'request': {
+                    'method': 'GET',
+                    'url': models_url,
+                    'headers': {
+                        'Authorization': f'Bearer {self._mask_api_key(api_key)}' if api_key else '',
+                    },
+                },
+                'response': payload,
+            },
+        }
+
     async def _req(
         self,
         query: pipeline_query.Query,

src/langbot/pkg/provider/modelmgr/requesters/ollamachat.py

@@ -8,6 +8,7 @@ import uuid
 import json
 
 import ollama
+import httpx
 
 from .. import errors, requester
 import langbot_plugin.api.entities.builtin.resource.tool as resource_tool
@@ -31,6 +32,60 @@ class OllamaChatCompletions(requester.ProviderAPIRequester):
         os.environ['OLLAMA_HOST'] = self.requester_cfg['base_url']
         self.client = ollama.AsyncClient(timeout=self.requester_cfg['timeout'])
 
+    def _infer_model_type(self, model_id: str) -> str:
+        normalized_model_id = (model_id or '').lower()
+        embedding_keywords = ('embedding', 'embed', 'bge-', 'e5-', 'm3e', 'gte-', 'text-embedding')
+        return 'embedding' if any(keyword in normalized_model_id for keyword in embedding_keywords) else 'llm'
+
+    def _infer_model_abilities(self, item: dict[str, typing.Any], model_id: str) -> list[str]:
+        normalized_model_id = (model_id or '').lower()
+        abilities: set[str] = set()
+        details = item.get('details', {}) or {}
+        families = details.get('families', []) or []
+        tokens = [normalized_model_id, str(details.get('family', '')).lower()]
+        tokens.extend(str(family).lower() for family in families)
+
+        if any(keyword in token for token in tokens for keyword in ('vision', 'vl', 'omni', 'llava', 'ocr')):
+            abilities.add('vision')
+        if any(keyword in token for token in tokens for keyword in ('tool', 'function')):
+            abilities.add('func_call')
+        return sorted(abilities)
+
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any]:
+        del api_key
+        models_url = f'{self.requester_cfg["base_url"].rstrip("/")}/api/tags'
+
+        async with httpx.AsyncClient(trust_env=True, timeout=self.requester_cfg['timeout']) as client:
+            response = await client.get(models_url)
+            response.raise_for_status()
+            payload = response.json()
+
+        models: list[dict[str, typing.Any]] = []
+        for item in payload.get('models', []):
+            model_id = item.get('model') or item.get('name')
+            if not model_id:
+                continue
+            models.append(
+                {
+                    'id': model_id,
+                    'name': item.get('name', model_id),
+                    'type': self._infer_model_type(model_id),
+                    'abilities': self._infer_model_abilities(item, model_id),
+                }
+            )
+
+        models.sort(key=lambda item: (item['type'] != 'llm', item['name'].lower()))
+        return {
+            'models': models,
+            'debug': {
+                'request': {
+                    'method': 'GET',
+                    'url': models_url,
+                },
+                'response': payload,
+            },
+        }
+
     async def _req(
         self,
         args: dict,
@@ -104,6 +159,21 @@ class OllamaChatCompletions(requester.ProviderAPIRequester):
 
         return ret_msg
 
+    async def _prepare_messages(
+        self,
+        messages: typing.List[provider_message.Message],
+    ) -> list[dict]:
+        """Prepare messages for Ollama API request."""
+        req_messages: list = []
+        for m in messages:
+            msg_dict: dict = m.dict(exclude_none=True)
+            content: Any = msg_dict.get('content')
+            if isinstance(content, list):
+                if all(isinstance(part, dict) and part.get('type') == 'text' for part in content):
+                    msg_dict['content'] = '\n'.join(part['text'] for part in content)
+            req_messages.append(msg_dict)
+        return req_messages
+
     async def invoke_llm(
         self,
         query: pipeline_query.Query,
@@ -113,14 +183,7 @@ class OllamaChatCompletions(requester.ProviderAPIRequester):
         extra_args: dict[str, typing.Any] = {},
         remove_think: bool = False,
     ) -> provider_message.Message:
-        req_messages: list = []
-        for m in messages:
-            msg_dict: dict = m.dict(exclude_none=True)
-            content: Any = msg_dict.get('content')
-            if isinstance(content, list):
-                if all(isinstance(part, dict) and part.get('type') == 'text' for part in content):
-                    msg_dict['content'] = '\n'.join(part['text'] for part in content)
-            req_messages.append(msg_dict)
+        req_messages = await self._prepare_messages(messages)
         try:
             return await self._closure(
                 query=query,
@@ -133,6 +196,109 @@ class OllamaChatCompletions(requester.ProviderAPIRequester):
         except asyncio.TimeoutError:
             raise errors.RequesterError('请求超时')
 
+    async def invoke_llm_stream(
+        self,
+        query: pipeline_query.Query,
+        model: requester.RuntimeLLMModel,
+        messages: typing.List[provider_message.Message],
+        funcs: typing.List[resource_tool.LLMTool] = None,
+        extra_args: dict[str, typing.Any] = {},
+        remove_think: bool = False,
+    ) -> provider_message.MessageChunk:
+        req_messages = await self._prepare_messages(messages)
+
+        try:
+            args = extra_args.copy()
+            args['model'] = model.model_entity.name
+
+            # Process messages for Ollama format
+            msgs: list[dict] = req_messages.copy()
+            for msg in msgs:
+                if 'content' in msg and isinstance(msg['content'], list):
+                    text_content: list = []
+                    image_urls: list = []
+                    for me in msg['content']:
+                        if me['type'] == 'text':
+                            text_content.append(me['text'])
+                        elif me['type'] == 'image_base64':
+                            image_urls.append(me['image_base64'])
+                    msg['content'] = '\n'.join(text_content)
+                    msg['images'] = [url.split(',')[1] for url in image_urls]
+                if 'tool_calls' in msg:
+                    for tool_call in msg['tool_calls']:
+                        tool_call['function']['arguments'] = json.loads(tool_call['function']['arguments'])
+            args['messages'] = msgs
+
+            args['tools'] = []
+            if funcs:
+                tools = await self.ap.tool_mgr.generate_tools_for_openai(funcs)
+                if tools:
+                    args['tools'] = tools
+
+            args['stream'] = True
+
+            chunk_idx = 0
+            thinking_started = False
+            thinking_ended = False
+            role = 'assistant'
+
+            async for chunk in await self.client.chat(**args):
+                message: ollama.Message = chunk.message
+                done = chunk.done
+
+                delta_content = message.content or ''
+                reasoning_content = getattr(message, 'thinking', '') or ''
+
+                # Handle reasoning/thinking content
+                if reasoning_content:
+                    if remove_think:
+                        chunk_idx += 1
+                        continue
+
+                    if not thinking_started:
+                        thinking_started = True
+                        delta_content = '<think>\n' + reasoning_content
+                    else:
+                        delta_content = reasoning_content
+                elif thinking_started and not thinking_ended and delta_content:
+                    thinking_ended = True
+                    delta_content = '\n</think>\n' + delta_content
+
+                # Handle tool calls
+                tool_calls_data = None
+                if message.tool_calls:
+                    tool_calls_data = []
+                    for tc in message.tool_calls:
+                        tool_calls_data.append(
+                            {
+                                'id': uuid.uuid4().hex,
+                                'type': 'function',
+                                'function': {
+                                    'name': tc.function.name,
+                                    'arguments': json.dumps(tc.function.arguments),
+                                },
+                            }
+                        )
+
+                # Skip empty first chunk
+                if chunk_idx == 0 and not delta_content and not reasoning_content and not tool_calls_data:
+                    chunk_idx += 1
+                    continue
+
+                chunk_data = {
+                    'role': role,
+                    'content': delta_content if delta_content else None,
+                    'tool_calls': tool_calls_data,
+                    'is_final': bool(done),
+                }
+                chunk_data = {k: v for k, v in chunk_data.items() if v is not None}
+
+                yield provider_message.MessageChunk(**chunk_data)
+                chunk_idx += 1
+
+        except asyncio.TimeoutError:
+            raise errors.RequesterError('请求超时')
+
     async def invoke_embedding(
         self,
         model: requester.RuntimeEmbeddingModel,

src/langbot/pkg/provider/modelmgr/requesters/openrouterchatcmpl.py

@@ -15,3 +15,11 @@ class OpenRouterChatCompletions(modelscopechatcmpl.ModelScopeChatCompletions):
         'base_url': 'https://openrouter.ai/api/v1',
         'timeout': 120,
     }
+
+    async def scan_models(self, api_key: str | None = None) -> dict[str, typing.Any]:
+        original_base_url = self.requester_cfg.get('base_url', '')
+        self.requester_cfg['base_url'] = 'https://openrouter.ai/api/v1'
+        try:
+            return await super().scan_models(api_key)
+        finally:
+            self.requester_cfg['base_url'] = original_base_url

src/langbot/pkg/provider/modelmgr/requesters/openrouterchatcmpl.yaml

@@ -25,6 +25,7 @@ spec:
   support_type:
   - llm
   - text-embedding
+  - rerank
   provider_category: maas
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/seekdb.svg

@@ -1,8 +1,17 @@
-<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
-  <rect width="24" height="24" rx="5" fill="#1E3A5F"/>
-  <path d="M6 12C6 8.68629 8.68629 6 12 6C15.3137 6 18 8.68629 18 12" stroke="#4FC3F7" stroke-width="2" stroke-linecap="round"/>
-  <path d="M18 12C18 15.3137 15.3137 18 12 18C8.68629 18 6 15.3137 6 12" stroke="#81D4FA" stroke-width="2" stroke-linecap="round"/>
-  <circle cx="12" cy="12" r="2" fill="#4FC3F7"/>
-  <circle cx="6" cy="12" r="1.5" fill="#81D4FA"/>
-  <circle cx="18" cy="12" r="1.5" fill="#4FC3F7"/>
-</svg>
+<svg id="_图层_1" data-name="图层 1" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 334.84 76.22">
+  <defs>
+    <style>
+      .cls-1 {
+        fill: currentColor;
+      }
+    </style>
+  </defs>
+  <path class="cls-1" d="M308.56,23.63c-5.04,0-9.73,1.43-13.73,3.88V1.08l-12.56,4.61v70h12.56v-3.35c4,2.46,8.71,3.88,13.73,3.88,14.49,0,26.29-11.79,26.29-26.29s-11.79-26.29-26.29-26.29h0ZM308.56,63.88c-6.87,0-12.57-4.98-13.73-11.51v-4.91c1.16-6.54,6.88-11.51,13.73-11.51,7.7,0,13.96,6.26,13.96,13.96s-6.26,13.96-13.96,13.96Z"></path>
+  <path class="cls-1" d="M255.54,5.69v21.83c-4-2.46-8.71-3.88-13.73-3.88-14.49,0-26.29,11.79-26.29,26.29s11.79,26.29,26.29,26.29c5.04,0,9.73-1.43,13.73-3.88v3.35h12.56V1.08l-12.56,4.61ZM241.81,63.88c-7.7,0-13.96-6.26-13.96-13.96s6.26-13.96,13.96-13.96c6.87,0,12.57,4.98,13.73,11.51v4.91c-1.16,6.54-6.88,11.51-13.73,11.51Z"></path>
+  <polygon class="cls-1" points="195.35 52.2 186.65 61.17 200.64 75.62 209.32 75.62 218.01 75.62 195.35 52.2"></polygon>
+  <path class="cls-1" d="M167.14,4.59c.65,3.99.68,8.04.03,12.15-.03.17.16.3.31.21,3.82-2.21,7.82-3.69,12.01-4.33.12-.02.19-.13.17-.23-.68-4.13-.61-8.18-.03-12.16.02-.17-.16-.3-.31-.2-4.01,2.31-8.01,3.81-12.01,4.34-.12.01-.19.12-.17.23h0Z"></path>
+  <path class="cls-1" d="M198.75,24.09l-19.07,19.72v-25.57c-4.49.67-8.7,2.11-12.56,4.57v52.83h12.56v-13.87l3.78-3.9.02.02,8.68-8.97-.02-.02,23.98-24.8h-17.37Z"></path>
+  <path class="cls-1" d="M145.03,57.86c-2.56,4.45-7.17,7.2-12.13,7.2-5.96,0-11.3-3.96-13.32-9.85h38.87l.08-.42c.29-1.5.42-3.06.42-4.65,0-14.37-11.69-26.06-26.06-26.06s-26.06,11.69-26.06,26.06,11.69,26.06,26.06,26.06c9.63,0,18.43-5.28,22.98-13.77l.26-.49-11.1-4.08h-.01ZM132.88,35.19h.03c5.96,0,11.3,3.96,13.32,9.85h-26.67c2.02-5.89,7.36-9.85,13.32-9.85Z"></path>
+  <path class="cls-1" d="M75.92,65.07c-5.96,0-11.29-3.96-13.32-9.85h38.87l.08-.42c.29-1.5.42-3.06.42-4.65,0-14.37-11.69-26.06-26.06-26.06s-26.06,11.69-26.06,26.06,11.69,26.06,26.06,26.06c9.63,0,18.43-5.28,22.98-13.77l.26-.49h0l-11.1-4.08c-2.56,4.45-7.17,7.2-12.13,7.2h-.01ZM75.92,35.19h.03c5.96,0,11.29,3.96,13.32,9.85h-26.67c2.03-5.89,7.36-9.85,13.32-9.85Z"></path>
+  <path class="cls-1" d="M30.43,45.58l-10.2-1.91c-3.03-.56-4.98-2.25-4.98-4.33,0-1.5,1.61-4.35,7.68-4.35,5.53,0,9.36,3.5,10.25,6.26l10.9-4-.14-.42c-1.17-3.54-3.5-6.58-6.94-9.04-3.49-2.49-8.04-3.69-13.88-3.69s-10.98,1.5-14.78,4.34c-3.88,2.91-5.84,6.76-5.84,11.46,0,7.98,4.72,12.77,14.42,14.64l9.9,1.81c3.05.61,4.94,2.27,4.94,4.33,0,2.61-3.58,4.44-8.7,4.44-5.79,0-9.9-3.72-11.85-7.14L0,62.1l.14.39c1.3,3.8,3.89,7.07,7.7,9.71,3.78,2.6,8.65,3.95,14.51,3.98l.25.03c6.87,0,12.55-1.57,16.43-4.53,3.98-3.05,6-6.99,6-11.74,0-3.73-1.14-6.7-3.6-9.33-2.27-2.42-5.98-4.11-10.98-5.02h-.02Z"></path>
+</svg>

src/langbot/pkg/provider/modelmgr/requesters/seekdbembed.py

@@ -46,14 +46,15 @@ class SeekDBEmbedding(requester.ProviderAPIRequester):
         extra_args: dict[str, typing.Any] = {},
     ) -> typing.List[typing.List[float]]:
         """Generate embeddings using SeekDB's built-in embedding function."""
+        if self._embedding_function is None:
+            await self.initialize()
+
         try:
-            if self._embedding_function is None:
-                await self.initialize()
-
-            if self._embedding_function is None:
-                raise RuntimeError('SeekDB embedding function initialization failed')
-
-            return self._embedding_function(input_text)
+            result = self._embedding_function(input_text)
+            # Ensure JSON serialization compatibility
+            if isinstance(result, list):
+                return [item.tolist() if hasattr(item, 'tolist') else item for item in result]
+            return result.tolist() if hasattr(result, 'tolist') else result
         except Exception as e:
             from .. import errors
 

src/langbot/pkg/provider/modelmgr/requesters/siliconflowchatcmpl.yaml

@@ -25,6 +25,7 @@ spec:
   support_type:
   - llm
   - text-embedding
+  - rerank
   provider_category: maas
 execution:
   python:

src/langbot/pkg/provider/modelmgr/requesters/voyageai.svg

@@ -0,0 +1 @@
+<svg height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>Voyage</title><path d="M5.407 0v.066a.974.974 0 00-.048.245c-.011.11-.016.208-.016.295 0 .339.043.715.128 1.13.097.405.274.912.531 1.524l7.125 16.366L20.011 3.39c.161-.404.333-.846.515-1.327.182-.48.273-.966.273-1.458a1.406 1.406 0 00-.096-.54V0H24v.066c-.204.207-.45.578-.74 1.114-.29.535-.606 1.195-.949 1.982L13.095 24h-1.287L3.075 3.965c-.204-.47-.418-.923-.644-1.36-.214-.437-.418-.83-.61-1.18-.194-.36-.365-.66-.515-.9A5.666 5.666 0 001 .064V0h4.407z" fill="#012E33"></path></svg>

src/langbot/pkg/provider/modelmgr/requesters/voyageairerank.yaml

@@ -0,0 +1,31 @@
+apiVersion: v1
+kind: LLMAPIRequester
+metadata:
+  name: voyageai-rerank
+  label:
+    en_US: Voyage AI
+    zh_Hans: Voyage AI
+  icon: voyageai.svg
+spec:
+  config:
+  - name: base_url
+    label:
+      en_US: Base URL
+      zh_Hans: 基础 URL
+    type: string
+    required: true
+    default: https://api.voyageai.com/v1
+  - name: timeout
+    label:
+      en_US: Timeout
+      zh_Hans: 超时时间
+    type: integer
+    required: true
+    default: 120
+  support_type:
+  - rerank
+  provider_category: manufacturer
+execution:
+  python:
+    path: ./chatcmpl.py
+    attr: OpenAIChatCompletions

src/langbot/pkg/provider/runners/dashscopeapi.py

@@ -107,7 +107,7 @@ class DashScopeAPIRunner(runner.RequestRunner):
 
         plain_text, image_ids = await self._preprocess_user_message(query)
         has_thoughts = True  # 获取思考过程
-        remove_think = self.pipeline_config['output'].get('misc', '').get('remove-think')
+        remove_think = self.pipeline_config['output'].get('misc', {}).get('remove-think')
         if remove_think:
             has_thoughts = False
         # 发送对话请求
@@ -141,7 +141,7 @@ class DashScopeAPIRunner(runner.RequestRunner):
                 idx_chunk += 1
                 # 获取流式传输的output
                 stream_output = chunk.get('output', {})
-                stream_think = stream_output.get('thoughts', [])
+                stream_think = stream_output.get('thoughts') or []
                 if stream_think and stream_think[0].get('thought'):
                     if not think_start:
                         think_start = True
@@ -149,7 +149,7 @@ class DashScopeAPIRunner(runner.RequestRunner):
                     else:
                         # 继续输出 reasoning_content
                         pending_content += stream_think[0].get('thought')
-                elif (not stream_think or stream_think[0].get('thought') == '') and not think_end:
+                elif think_start and (not stream_think or stream_think[0].get('thought') == '') and not think_end:
                     think_end = True
                     pending_content += '\n</think>\n'
                 if stream_output.get('text') is not None:
@@ -188,15 +188,15 @@ class DashScopeAPIRunner(runner.RequestRunner):
                 idx_chunk += 1
                 # 获取流式传输的output
                 stream_output = chunk.get('output', {})
-                stream_think = stream_output.get('thoughts', [])
-                if stream_think[0].get('thought'):
+                stream_think = stream_output.get('thoughts') or []
+                if stream_think and stream_think[0].get('thought'):
                     if not think_start:
                         think_start = True
                         pending_content += f'<think>\n{stream_think[0].get("thought")}'
                     else:
                         # 继续输出 reasoning_content
                         pending_content += stream_think[0].get('thought')
-                elif stream_think[0].get('thought') == '' and not think_end:
+                elif think_start and (not stream_think or stream_think[0].get('thought') == '') and not think_end:
                     think_end = True
                     pending_content += '\n</think>\n'
                 if stream_output.get('text') is not None:

src/langbot/pkg/provider/runners/localagent.py

@@ -172,6 +172,45 @@ class LocalAgentRunner(runner.RequestRunner):
                 if result:
                     all_results.extend(result)
 
+            # Rerank step: re-score results using a rerank model if configured
+            local_agent_config = query.pipeline_config.get('ai', {}).get('local-agent', {})
+            rerank_model_uuid = local_agent_config.get('rerank-model', '')
+            if rerank_model_uuid == '__none__':
+                rerank_model_uuid = ''
+            self.ap.logger.info(
+                f'Rerank config: model_uuid={rerank_model_uuid!r}, '
+                f'results={len(all_results)}, '
+                f'local_agent_keys={list(local_agent_config.keys())}'
+            )
+            if all_results and rerank_model_uuid:
+                try:
+                    rerank_model = await self.ap.model_mgr.get_rerank_model_by_uuid(rerank_model_uuid)
+                    rerank_top_k = int(local_agent_config.get('rerank-top-k', 5))
+
+                    doc_texts = []
+                    for entry in all_results:
+                        text = ' '.join(c.text for c in entry.content if c.type == 'text' and c.text)
+                        doc_texts.append(text)
+
+                    doc_texts_capped = doc_texts[:64]
+                    scores = await rerank_model.provider.invoke_rerank(
+                        model=rerank_model,
+                        query=user_message_text,
+                        documents=doc_texts_capped,
+                    )
+
+                    scored = sorted(scores, key=lambda x: x.get('relevance_score', 0), reverse=True)
+                    top_indices = [s['index'] for s in scored[:rerank_top_k] if s['index'] < len(all_results)]
+                    all_results = [all_results[i] for i in top_indices]
+
+                    self.ap.logger.info(
+                        f'Rerank complete: {len(doc_texts)} docs reranked -> top {len(all_results)} kept (top_k={rerank_top_k})'
+                    )
+                except ValueError:
+                    self.ap.logger.warning(f'Rerank model {rerank_model_uuid} not found, skipping rerank')
+                except Exception as e:
+                    self.ap.logger.warning(f'Rerank failed, using original order: {e}')
+
             final_user_message_text = ''
 
             if all_results:

src/langbot/pkg/provider/runners/n8nsvapi.py

@@ -70,11 +70,12 @@ class N8nServiceAPIRunner(runner.RequestRunner):
 
         return plain_text
 
-    async def _process_stream_response(
+    async def _process_response(
         self, response: aiohttp.ClientResponse
     ) -> typing.AsyncGenerator[provider_message.Message, None]:
-        """处理流式响应——支持部分 JSON 和多个 JSON 对象在同一 chunk 的情况"""
+        """处理响应——支持流式格式和普通 JSON 格式"""
         full_content = ''
+        full_text = ''
         chunk_idx = 0
         is_final = False
         message_idx = 0
@@ -93,6 +94,7 @@ class N8nServiceAPIRunner(runner.RequestRunner):
                 else:
                     chunk_str = str(raw_chunk)
 
+                full_text += chunk_str
                 buffer += chunk_str
 
                 # 尝试从 buffer 中循环解析出 JSON 对象（处理多个对象或部分对象）
@@ -115,7 +117,7 @@ class N8nServiceAPIRunner(runner.RequestRunner):
                         elif obj.get('type') == 'end':
                             is_final = True
 
-                        if is_final or chunk_idx % 8 == 0:
+                        if is_final or (chunk_idx > 0 and chunk_idx % 8 == 0):
                             message_idx += 1
                             yield provider_message.MessageChunk(
                                 role='assistant',
@@ -142,6 +144,7 @@ class N8nServiceAPIRunner(runner.RequestRunner):
                     obj, _ = decoder.raw_decode(buffer)
                     if isinstance(obj, dict):
                         if obj.get('type') == 'item' and 'content' in obj:
+                            chunk_idx += 1
                             full_content += obj['content']
                         elif obj.get('type') == 'end':
                             is_final = True
@@ -156,6 +159,28 @@ class N8nServiceAPIRunner(runner.RequestRunner):
                 preview = buffer[:200]
                 self.ap.logger.warning(f'Failed to parse remaining buffer: {e}; buffer preview: {preview}')
 
+        # n8n 返回普通 JSON 格式（无任何流式 type:item 内容）
+        if chunk_idx == 0:
+            output_content = ''
+            try:
+                response_data = json.loads(full_text.strip())
+                if isinstance(response_data, dict):
+                    if self.output_key in response_data:
+                        output_content = response_data[self.output_key]
+                    else:
+                        output_content = json.dumps(response_data, ensure_ascii=False)
+                else:
+                    output_content = full_text
+            except json.JSONDecodeError:
+                output_content = full_text
+            self.ap.logger.debug(f'n8n webhook response (non-stream): {full_text[:200]}')
+            yield provider_message.MessageChunk(
+                role='assistant',
+                content=output_content,
+                is_final=True,
+                msg_sequence=message_idx + 1,
+            )
+
     async def _call_webhook(self, query: pipeline_query.Query) -> typing.AsyncGenerator[provider_message.Message, None]:
         """调用n8n webhook"""
         # 生成会话ID（如果不存在）
@@ -220,49 +245,22 @@ class N8nServiceAPIRunner(runner.RequestRunner):
 
             # 调用webhook
             session = httpclient.get_session()
-            if is_stream:
-                # 流式请求
-                async with session.post(
-                    self.webhook_url, json=payload, headers=headers, auth=auth, timeout=self.timeout
-                ) as response:
-                    if response.status != 200:
-                        error_text = await response.text()
-                        self.ap.logger.error(f'n8n webhook call failed: {response.status}, {error_text}')
-                        raise Exception(f'n8n webhook call failed: {response.status}, {error_text}')
+            async with session.post(
+                self.webhook_url, json=payload, headers=headers, auth=auth, timeout=self.timeout
+            ) as response:
+                if response.status != 200:
+                    error_text = await response.text()
+                    self.ap.logger.error(f'n8n webhook call failed: {response.status}, {error_text}')
+                    raise Exception(f'n8n webhook call failed: {response.status}, {error_text}')
 
-                    # 处理流式响应
-                    async for chunk in self._process_stream_response(response):
+                async for chunk in self._process_response(response):
+                    if is_stream:
                         yield chunk
-            else:
-                async with session.post(
-                    self.webhook_url, json=payload, headers=headers, auth=auth, timeout=self.timeout
-                ) as response:
-                    try:
-                        async for chunk in self._process_stream_response(response):
-                            output_content = chunk.content if chunk.is_final else ''
-                    except:
-                        # 非流式请求（保持原有逻辑）
-                        if response.status != 200:
-                            error_text = await response.text()
-                            self.ap.logger.error(f'n8n webhook call failed: {response.status}, {error_text}')
-                            raise Exception(f'n8n webhook call failed: {response.status}, {error_text}')
-
-                        # 解析响应
-                        response_data = await response.json()
-                        self.ap.logger.debug(f'n8n webhook response: {response_data}')
-
-                        # 从响应中提取输出
-                        if self.output_key in response_data:
-                            output_content = response_data[self.output_key]
-                        else:
-                            # 如果没有指定的输出键，则使用整个响应
-                            output_content = json.dumps(response_data, ensure_ascii=False)
-
-                    # 返回消息
-                    yield provider_message.Message(
-                        role='assistant',
-                        content=output_content,
-                    )
+                    elif chunk.is_final:
+                        yield provider_message.Message(
+                            role='assistant',
+                            content=chunk.content,
+                        )
         except Exception as e:
             self.ap.logger.error(f'n8n webhook call exception: {str(e)}')
             raise N8nAPIError(f'n8n webhook call exception: {str(e)}')

src/langbot/pkg/telemetry/telemetry.py

@@ -60,7 +60,16 @@ class TelemetryManager:
                     except Exception:
                         sanitized['query_id'] = str(sanitized.get('query_id', ''))
 
-                for sfield in ('adapter', 'runner', 'runner_category', 'model_name', 'version', 'error', 'timestamp'):
+                for sfield in (
+                    'adapter',
+                    'runner',
+                    'runner_category',
+                    'model_name',
+                    'version',
+                    'edition',
+                    'error',
+                    'timestamp',
+                ):
                     v = sanitized.get(sfield)
                     sanitized[sfield] = '' if v is None else str(v)
 

src/langbot/pkg/utils/constants.py

@@ -2,7 +2,7 @@ import langbot
 
 semantic_version = f'v{langbot.__version__}'
 
-required_database_version = 24
+required_database_version = 25
 """Tag the version of the database schema, used to check if the database needs to be migrated"""
 
 debug_mode = False

src/langbot/pkg/utils/paths.py

@@ -38,28 +38,31 @@ def get_frontend_path() -> str:
     """
     Get the path to the frontend build files.
 
-    Returns the path to web/out directory, handling both:
+    Returns the path to web/dist directory (Vite build output), handling both:
     - Development mode: running from source directory
     - Package mode: installed via pip/uvx
+    - Legacy mode: web/out (Next.js, for backward compatibility)
     """
-    # First, check if we're running from source directory
-    if _check_if_source_install() and os.path.exists('web/out'):
-        return 'web/out'
+    # Check both dist (Vite) and out (legacy Next.js) paths
+    for dirname in ('dist', 'out'):
+        web_dir = f'web/{dirname}'
 
-    # Second, check current directory for web/out (in case user is in source dir)
-    if os.path.exists('web/out'):
-        return 'web/out'
+        # First, check if we're running from source directory
+        if _check_if_source_install() and os.path.exists(web_dir):
+            return web_dir
 
-    # Third, find it relative to the package installation
-    # Get the directory where this file is located
-    # paths.py is in pkg/utils/, so parent.parent goes up to pkg/, then parent again goes up to the package root
-    pkg_dir = Path(__file__).parent.parent.parent
-    frontend_path = pkg_dir / 'web' / 'out'
-    if frontend_path.exists():
-        return str(frontend_path)
+        # Second, check current directory
+        if os.path.exists(web_dir):
+            return web_dir
+
+        # Third, find it relative to the package installation
+        pkg_dir = Path(__file__).parent.parent.parent
+        frontend_path = pkg_dir / 'web' / dirname
+        if frontend_path.exists():
+            return str(frontend_path)
 
     # Return the default path (will be checked by caller)
-    return 'web/out'
+    return 'web/dist'
 
 
 def get_resource_path(resource: str) -> str:

src/langbot/pkg/vector/mgr.py

@@ -2,11 +2,6 @@ from __future__ import annotations
 
 from ..core import app
 from .vdb import VectorDatabase, SearchType
-from .vdbs.chroma import ChromaVectorDatabase
-from .vdbs.qdrant import QdrantVectorDatabase
-from .vdbs.seekdb import SeekDBVectorDatabase
-from .vdbs.milvus import MilvusVectorDatabase
-from .vdbs.pgvector_db import PgVectorDatabase
 
 
 class VectorDBManager:
@@ -22,17 +17,25 @@ class VectorDBManager:
             vdb_type = kb_config.get('use')
 
             if vdb_type == 'chroma':
+                from .vdbs.chroma import ChromaVectorDatabase
+
                 self.vector_db = ChromaVectorDatabase(self.ap)
                 self.ap.logger.info('Initialized Chroma vector database backend.')
 
             elif vdb_type == 'qdrant':
+                from .vdbs.qdrant import QdrantVectorDatabase
+
                 self.vector_db = QdrantVectorDatabase(self.ap)
                 self.ap.logger.info('Initialized Qdrant vector database backend.')
             elif vdb_type == 'seekdb':
+                from .vdbs.seekdb import SeekDBVectorDatabase
+
                 self.vector_db = SeekDBVectorDatabase(self.ap)
                 self.ap.logger.info('Initialized SeekDB vector database backend.')
 
             elif vdb_type == 'milvus':
+                from .vdbs.milvus import MilvusVectorDatabase
+
                 # Get Milvus configuration
                 milvus_config = kb_config.get('milvus', {})
                 uri = milvus_config.get('uri', './data/milvus.db')
@@ -42,6 +45,8 @@ class VectorDBManager:
                 self.ap.logger.info('Initialized Milvus vector database backend.')
 
             elif vdb_type == 'pgvector':
+                from .vdbs.pgvector_db import PgVectorDatabase
+
                 # Get pgvector configuration
                 pgvector_config = kb_config.get('pgvector', {})
                 connection_string = pgvector_config.get('connection_string')
@@ -60,9 +65,13 @@ class VectorDBManager:
                 self.ap.logger.info('Initialized pgvector database backend.')
 
             else:
+                from .vdbs.chroma import ChromaVectorDatabase
+
                 self.vector_db = ChromaVectorDatabase(self.ap)
                 self.ap.logger.warning('No valid vector database backend configured, defaulting to Chroma.')
         else:
+            from .vdbs.chroma import ChromaVectorDatabase
+
             self.vector_db = ChromaVectorDatabase(self.ap)
             self.ap.logger.warning('No vector database backend configured, defaulting to Chroma.')
 

src/langbot/pkg/vector/vdbs/__init__.py

@@ -1,7 +1 @@
 """Vector database implementations for LangBot."""
-
-from .chroma import ChromaVectorDatabase
-from .qdrant import QdrantVectorDatabase
-from .seekdb import SeekDBVectorDatabase
-
-__all__ = ['ChromaVectorDatabase', 'QdrantVectorDatabase', 'SeekDBVectorDatabase']

src/langbot/templates/config.yaml

@@ -20,6 +20,7 @@ system:
     edition: community
     recovery_key: ''
     allow_modify_login_info: true
+    disabled_adapters: []
     limitation:
         max_bots: -1
         max_pipelines: -1

src/langbot/templates/default-pipeline-config.json

@@ -52,7 +52,9 @@
                     "content": "You are a helpful assistant."
                 }
             ],
-            "knowledge-bases": []
+            "knowledge-bases": [],
+            "rerank-model": "",
+            "rerank-top-k": 5
         },
         "dify-service-api": {
             "base-url": "https://api.dify.ai/v1",

src/langbot/templates/metadata/pipeline/ai.yaml

@@ -104,6 +104,34 @@ stages:
           field: __system.is_wizard
           operator: neq
           value: true
+      - name: rerank-model
+        label:
+          en_US: Rerank Model
+          zh_Hans: 重排序模型
+        description:
+          en_US: Optional rerank model to improve retrieval quality by re-scoring retrieved chunks
+          zh_Hans: 可选的重排序模型，通过重新评分检索结果来提升检索质量
+        type: rerank-model-selector
+        required: false
+        default: ''
+        show_if:
+          field: knowledge-bases
+          operator: neq
+          value: []
+      - name: rerank-top-k
+        label:
+          en_US: Rerank Top K
+          zh_Hans: 重排序保留数量
+        description:
+          en_US: Number of top results to keep after reranking
+          zh_Hans: 重排序后保留的最相关结果数量
+        type: integer
+        required: false
+        default: 5
+        show_if:
+          field: rerank-model
+          operator: neq
+          value: ''
   - name: dify-service-api
     label:
       en_US: Dify Service API

tests/unit_tests/pipeline/test_n8nsvapi.py

@@ -0,0 +1,328 @@
+"""
+Unit tests for N8nServiceAPIRunner._process_response
+
+Tests cover four scenarios:
+- Stream adapter + n8n stream format  (type:item/end)
+- Stream adapter + n8n plain JSON
+- Non-stream adapter + n8n stream format
+- Non-stream adapter + n8n plain JSON
+"""
+
+from __future__ import annotations
+
+import json
+import sys
+from unittest.mock import AsyncMock, MagicMock, Mock, patch
+
+# Break the circular import chain before importing n8nsvapi:
+#   n8nsvapi → runner → app → pipelinemgr → all runners → runner (partially init)
+_mock_runner = MagicMock()
+_mock_runner.runner_class = lambda name: (lambda cls: cls)  # no-op decorator
+_mock_runner.RequestRunner = object
+sys.modules.setdefault('langbot.pkg.provider.runner', _mock_runner)
+sys.modules.setdefault('langbot.pkg.core.app', MagicMock())
+sys.modules.setdefault('langbot.pkg.utils.httpclient', MagicMock())
+
+import pytest
+import langbot_plugin.api.entities.builtin.provider.message as provider_message
+from langbot.pkg.provider.runners.n8nsvapi import N8nServiceAPIRunner
+
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+def make_runner(output_key: str = 'response') -> N8nServiceAPIRunner:
+    ap = Mock()
+    ap.logger = Mock()
+    pipeline_config = {
+        'ai': {
+            'n8n-service-api': {
+                'webhook-url': 'http://test-n8n/webhook',
+                'output-key': output_key,
+                'auth-type': 'none',
+            }
+        }
+    }
+    return N8nServiceAPIRunner(ap, pipeline_config)
+
+
+def make_mock_response(chunks: list[bytes | str], status: int = 200):
+    """Build a minimal aiohttp.ClientResponse mock with iter_chunked support."""
+    response = Mock()
+    response.status = status
+
+    async def iter_chunked(size):
+        for chunk in chunks:
+            yield chunk
+
+    response.content = Mock()
+    response.content.iter_chunked = iter_chunked
+    return response
+
+
+async def collect_chunks(runner: N8nServiceAPIRunner, chunks: list[bytes | str]):
+    """Run _process_response and collect all yielded MessageChunks."""
+    response = make_mock_response(chunks)
+    result = []
+    async for chunk in runner._process_response(response):
+        result.append(chunk)
+    return result
+
+
+# ---------------------------------------------------------------------------
+# _process_response: stream format (type:item/end)
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_stream_format_single_item():
+    """Single item + end in one chunk yields final chunk with full content."""
+    runner = make_runner()
+    data = b'{"type":"item","content":"hello"}{"type":"end"}'
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) >= 1
+    final = chunks[-1]
+    assert final.is_final is True
+    assert final.content == 'hello'
+
+
+@pytest.mark.asyncio
+async def test_stream_format_multi_item_accumulates():
+    """Multiple items accumulate into full_content."""
+    runner = make_runner()
+    chunks_data = [
+        b'{"type":"item","content":"foo"}',
+        b'{"type":"item","content":"bar"}',
+        b'{"type":"end"}',
+    ]
+
+    chunks = await collect_chunks(runner, chunks_data)
+
+    final = chunks[-1]
+    assert final.is_final is True
+    assert final.content == 'foobar'
+
+
+@pytest.mark.asyncio
+async def test_stream_format_batches_every_8_items():
+    """Every 8th item triggers an intermediate yield before the final."""
+    runner = make_runner()
+    items = [f'{{"type":"item","content":"{i}"}}' for i in range(8)]
+    items.append('{"type":"end"}')
+    data = ''.join(items).encode()
+
+    chunks = await collect_chunks(runner, [data])
+
+    # At least the batch yield at chunk_idx==8 + final yield
+    assert len(chunks) >= 2
+    assert chunks[-1].is_final is True
+
+
+@pytest.mark.asyncio
+async def test_stream_format_split_across_network_chunks():
+    """JSON split across multiple network chunks is reassembled correctly."""
+    runner = make_runner()
+    part1 = b'{"type":"item","con'
+    part2 = b'tent":"world"}{"type":"end"}'
+
+    chunks = await collect_chunks(runner, [part1, part2])
+
+    final = chunks[-1]
+    assert final.is_final is True
+    assert final.content == 'world'
+
+
+@pytest.mark.asyncio
+async def test_stream_format_no_spurious_empty_yield():
+    """chunk_idx==0 guard prevents spurious empty yield before any item is received."""
+    runner = make_runner()
+    # Send some non-stream JSON first, then stream
+    data = b'{"type":"item","content":"x"}{"type":"end"}'
+
+    chunks = await collect_chunks(runner, [data])
+
+    # No chunk should have empty content before the real content arrives
+    non_final = [c for c in chunks if not c.is_final]
+    for c in non_final:
+        assert c.content  # must be non-empty
+
+
+# ---------------------------------------------------------------------------
+# _process_response: plain JSON fallback
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+async def test_plain_json_with_output_key():
+    """Plain JSON with matching output_key extracts value via output_key."""
+    runner = make_runner(output_key='response')
+    data = json.dumps({'response': 'hello world'}).encode()
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) == 1
+    assert chunks[0].is_final is True
+    assert chunks[0].content == 'hello world'
+
+
+@pytest.mark.asyncio
+async def test_plain_json_output_key_not_found():
+    """Plain JSON without output_key falls back to entire JSON string."""
+    runner = make_runner(output_key='response')
+    payload = {'other_key': 'hello'}
+    data = json.dumps(payload).encode()
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) == 1
+    assert chunks[0].is_final is True
+    assert json.loads(chunks[0].content) == payload
+
+
+@pytest.mark.asyncio
+async def test_plain_json_output_key_empty_string():
+    """output_key present but value is empty string — returns empty string, not whole JSON."""
+    runner = make_runner(output_key='response')
+    data = json.dumps({'response': ''}).encode()
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) == 1
+    assert chunks[0].is_final is True
+    assert chunks[0].content == ''
+
+
+@pytest.mark.asyncio
+async def test_plain_json_non_dict_response():
+    """Plain JSON array falls back to raw text."""
+    runner = make_runner()
+    data = b'["a", "b"]'
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) == 1
+    assert chunks[0].is_final is True
+    assert chunks[0].content == '["a", "b"]'
+
+
+@pytest.mark.asyncio
+async def test_invalid_json_returns_raw_text():
+    """Non-JSON response returns raw text as-is."""
+    runner = make_runner()
+    data = b'plain text response'
+
+    chunks = await collect_chunks(runner, [data])
+
+    assert len(chunks) == 1
+    assert chunks[0].is_final is True
+    assert chunks[0].content == 'plain text response'
+
+
+# ---------------------------------------------------------------------------
+# _call_webhook: output type depends on is_stream
+# ---------------------------------------------------------------------------
+
+def make_query(is_stream: bool):
+    """Build a minimal Query mock."""
+    query = Mock()
+    query.adapter = AsyncMock()
+    query.adapter.is_stream_output_supported = AsyncMock(return_value=is_stream)
+
+    session = Mock()
+    session.using_conversation = Mock()
+    session.using_conversation.uuid = 'test-uuid'
+    session.launcher_type = Mock()
+    session.launcher_type.value = 'person'
+    session.launcher_id = '12345'
+    query.session = session
+
+    query.user_message = Mock()
+    query.user_message.content = 'hi'
+    query.variables = {}
+    return query
+
+
+def make_http_session_mock(response_bytes: bytes, status: int = 200):
+    """Mock httpclient.get_session() returning a session whose post() yields response_bytes."""
+    mock_response = make_mock_response([response_bytes], status=status)
+    mock_response.status = status
+
+    mock_cm = AsyncMock()
+    mock_cm.__aenter__ = AsyncMock(return_value=mock_response)
+    mock_cm.__aexit__ = AsyncMock(return_value=False)
+
+    mock_session = Mock()
+    mock_session.post = Mock(return_value=mock_cm)
+    return mock_session
+
+
+@pytest.mark.asyncio
+async def test_call_webhook_nonstream_adapter_plain_json():
+    """Non-stream adapter + plain JSON → single Message with output_key value."""
+    runner = make_runner(output_key='response')
+    query = make_query(is_stream=False)
+    http_session = make_http_session_mock(json.dumps({'response': 'result text'}).encode())
+
+    with patch('langbot.pkg.provider.runners.n8nsvapi.httpclient.get_session', return_value=http_session):
+        results = []
+        async for msg in runner._call_webhook(query):
+            results.append(msg)
+
+    assert len(results) == 1
+    assert isinstance(results[0], provider_message.Message)
+    assert results[0].content == 'result text'
+
+
+@pytest.mark.asyncio
+async def test_call_webhook_stream_adapter_stream_format():
+    """Stream adapter + stream format → MessageChunks, last is_final."""
+    runner = make_runner()
+    query = make_query(is_stream=True)
+    data = b'{"type":"item","content":"hi"}{"type":"end"}'
+    http_session = make_http_session_mock(data)
+
+    with patch('langbot.pkg.provider.runners.n8nsvapi.httpclient.get_session', return_value=http_session):
+        results = []
+        async for msg in runner._call_webhook(query):
+            results.append(msg)
+
+    assert all(isinstance(r, provider_message.MessageChunk) for r in results)
+    assert results[-1].is_final is True
+    assert results[-1].content == 'hi'
+
+
+@pytest.mark.asyncio
+async def test_call_webhook_stream_adapter_plain_json():
+    """Stream adapter + plain JSON → single MessageChunk with is_final=True."""
+    runner = make_runner(output_key='response')
+    query = make_query(is_stream=True)
+    data = json.dumps({'response': 'fallback'}).encode()
+    http_session = make_http_session_mock(data)
+
+    with patch('langbot.pkg.provider.runners.n8nsvapi.httpclient.get_session', return_value=http_session):
+        results = []
+        async for msg in runner._call_webhook(query):
+            results.append(msg)
+
+    assert all(isinstance(r, provider_message.MessageChunk) for r in results)
+    assert results[-1].is_final is True
+    assert results[-1].content == 'fallback'
+
+
+@pytest.mark.asyncio
+async def test_call_webhook_nonstream_adapter_stream_format():
+    """Non-stream adapter + stream format → single Message with accumulated content."""
+    runner = make_runner()
+    query = make_query(is_stream=False)
+    data = b'{"type":"item","content":"foo"}{"type":"item","content":"bar"}{"type":"end"}'
+    http_session = make_http_session_mock(data)
+
+    with patch('langbot.pkg.provider.runners.n8nsvapi.httpclient.get_session', return_value=http_session):
+        results = []
+        async for msg in runner._call_webhook(query):
+            results.append(msg)
+
+    assert len(results) == 1
+    assert isinstance(results[0], provider_message.Message)
+    assert results[0].content == 'foobar'

tests/unit_tests/platform/test_routing_rules.py

@@ -0,0 +1,280 @@
+"""
+RuntimeBot.resolve_pipeline_uuid and _match_operator unit tests
+"""
+
+from unittest.mock import Mock
+
+
+class TestMatchOperator:
+    """Test the _match_operator static method."""
+
+    @staticmethod
+    def _get_class():
+        from langbot.pkg.platform.botmgr import RuntimeBot
+
+        return RuntimeBot
+
+    def test_eq(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello', 'eq', 'hello') is True
+        assert cls._match_operator('hello', 'eq', 'world') is False
+
+    def test_neq(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello', 'neq', 'world') is True
+        assert cls._match_operator('hello', 'neq', 'hello') is False
+
+    def test_contains(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello world', 'contains', 'world') is True
+        assert cls._match_operator('hello world', 'contains', 'xyz') is False
+
+    def test_not_contains(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello world', 'not_contains', 'xyz') is True
+        assert cls._match_operator('hello world', 'not_contains', 'world') is False
+
+    def test_starts_with(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello world', 'starts_with', 'hello') is True
+        assert cls._match_operator('hello world', 'starts_with', 'world') is False
+
+    def test_regex(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello123', 'regex', r'\d+') is True
+        assert cls._match_operator('hello', 'regex', r'\d+') is False
+
+    def test_regex_invalid_pattern(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello', 'regex', r'[invalid') is False
+
+    def test_unknown_operator(self):
+        cls = self._get_class()
+        assert cls._match_operator('hello', 'unknown_op', 'hello') is False
+
+
+class TestResolvePipelineUuid:
+    """Test the resolve_pipeline_uuid method."""
+
+    @staticmethod
+    def _make_bot(default_pipeline: str, rules: list):
+        from langbot.pkg.platform.botmgr import RuntimeBot
+
+        bot_entity = Mock()
+        bot_entity.use_pipeline_uuid = default_pipeline
+        bot_entity.pipeline_routing_rules = rules
+
+        bot = object.__new__(RuntimeBot)
+        bot.bot_entity = bot_entity
+        return bot
+
+    def test_no_rules_returns_default(self):
+        bot = self._make_bot('default-uuid', [])
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_none_rules_returns_default(self):
+        bot = self._make_bot('default-uuid', None)
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_launcher_type_match(self):
+        rules = [
+            {
+                'type': 'launcher_type',
+                'operator': 'eq',
+                'value': 'group',
+                'pipeline_uuid': 'group-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('group', '123', 'hi')
+        assert uuid == 'group-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_launcher_id_match(self):
+        rules = [
+            {
+                'type': 'launcher_id',
+                'operator': 'eq',
+                'value': '12345',
+                'pipeline_uuid': 'vip-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '12345', 'hi')
+        assert uuid == 'vip-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '99999', 'hi')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_message_content_contains(self):
+        rules = [
+            {
+                'type': 'message_content',
+                'operator': 'contains',
+                'value': '紧急',
+                'pipeline_uuid': 'urgent-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', '这是紧急消息')
+        assert uuid == 'urgent-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', '普通消息')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_message_content_regex(self):
+        rules = [
+            {
+                'type': 'message_content',
+                'operator': 'regex',
+                'value': r'^/admin\b',
+                'pipeline_uuid': 'admin-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', '/admin help')
+        assert uuid == 'admin-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hello /admin')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_message_has_element_eq(self):
+        rules = [
+            {
+                'type': 'message_has_element',
+                'operator': 'eq',
+                'value': 'Image',
+                'pipeline_uuid': 'image-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi', ['Plain', 'Image'])
+        assert uuid == 'image-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi', ['Plain'])
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_message_has_element_neq(self):
+        rules = [
+            {
+                'type': 'message_has_element',
+                'operator': 'neq',
+                'value': 'Image',
+                'pipeline_uuid': 'text-only-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi', ['Plain'])
+        assert uuid == 'text-only-pipeline'
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi', ['Plain', 'Image'])
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_message_has_element_no_types_provided(self):
+        """When element types are not provided, should not match."""
+        rules = [
+            {
+                'type': 'message_has_element',
+                'operator': 'eq',
+                'value': 'Image',
+                'pipeline_uuid': 'image-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'default-uuid'
+        assert routed is False
+
+    def test_first_match_wins(self):
+        rules = [
+            {
+                'type': 'launcher_type',
+                'operator': 'eq',
+                'value': 'group',
+                'pipeline_uuid': 'first-pipeline',
+            },
+            {
+                'type': 'launcher_type',
+                'operator': 'eq',
+                'value': 'group',
+                'pipeline_uuid': 'second-pipeline',
+            },
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('group', '123', 'hi')
+        assert uuid == 'first-pipeline'
+        assert routed is True
+
+    def test_skip_invalid_rules(self):
+        rules = [
+            {'type': '', 'operator': 'eq', 'value': 'x', 'pipeline_uuid': 'p1'},
+            {'type': 'launcher_type', 'operator': 'eq', 'value': 'person', 'pipeline_uuid': ''},
+            {'type': 'launcher_type', 'operator': 'eq', 'value': 'person', 'pipeline_uuid': 'valid'},
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'valid'
+        assert routed is True
+
+    def test_default_operator_is_eq(self):
+        rules = [
+            {
+                'type': 'launcher_type',
+                'value': 'person',
+                'pipeline_uuid': 'person-pipeline',
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'hi')
+        assert uuid == 'person-pipeline'
+        assert routed is True
+
+    def test_discard_pipeline(self):
+        """When pipeline_uuid is __discard__, the message should be discarded."""
+        from langbot.pkg.platform.botmgr import RuntimeBot
+
+        rules = [
+            {
+                'type': 'message_content',
+                'operator': 'contains',
+                'value': 'spam',
+                'pipeline_uuid': RuntimeBot.PIPELINE_DISCARD,
+            }
+        ]
+        bot = self._make_bot('default-uuid', rules)
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'this is spam')
+        assert uuid == RuntimeBot.PIPELINE_DISCARD
+        assert routed is True
+
+        uuid, routed = bot.resolve_pipeline_uuid('person', '123', 'normal message')
+        assert uuid == 'default-uuid'
+        assert routed is False

uv.lock

@@ -186,6 +186,20 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/00/b7/e3bf5133d697a08128598c8d0abc5e16377b51465a33756de24fa7dee953/aiosqlite-0.22.1-py3-none-any.whl", hash = "sha256:21c002eb13823fad740196c5a2e9d8e62f6243bd9e7e4a1f87fb5e44ecb4fceb", size = 17405, upload-time = "2025-12-23T19:25:42.139Z" },
 ]
 
+[[package]]
+name = "alembic"
+version = "1.18.4"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "mako" },
+    { name = "sqlalchemy" },
+    { name = "typing-extensions" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/94/13/8b084e0f2efb0275a1d534838844926f798bd766566b1375174e2448cd31/alembic-1.18.4.tar.gz", hash = "sha256:cb6e1fd84b6174ab8dbb2329f86d631ba9559dd78df550b57804d607672cedbc", size = 2056725, upload-time = "2026-02-10T16:00:47.195Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/d2/29/6533c317b74f707ea28f8d633734dbda2119bbadfc61b2f3640ba835d0f7/alembic-1.18.4-py3-none-any.whl", hash = "sha256:a5ed4adcf6d8a4cb575f3d759f071b03cd6e5c7618eb796cb52497be25bfe19a", size = 263893, upload-time = "2026-02-10T16:00:49.997Z" },
+]
+
 [[package]]
 name = "annotated-types"
 version = "0.7.0"
@@ -1832,7 +1846,7 @@ wheels = [
 
 [[package]]
 name = "langbot"
-version = "4.9.5"
+version = "4.9.6"
 source = { editable = "." }
 dependencies = [
     { name = "aiocqhttp" },
@@ -1840,6 +1854,7 @@ dependencies = [
     { name = "aiohttp" },
     { name = "aioshutil" },
     { name = "aiosqlite" },
+    { name = "alembic" },
     { name = "anthropic" },
     { name = "argon2-cffi" },
     { name = "async-lru" },
@@ -1919,6 +1934,7 @@ requires-dist = [
     { name = "aiohttp", specifier = ">=3.11.18" },
     { name = "aioshutil", specifier = ">=1.5" },
     { name = "aiosqlite", specifier = ">=0.21.0" },
+    { name = "alembic", specifier = ">=1.15.0" },
     { name = "anthropic", specifier = ">=0.51.0" },
     { name = "argon2-cffi", specifier = ">=23.1.0" },
     { name = "async-lru", specifier = ">=2.0.5" },
@@ -1937,7 +1953,7 @@ requires-dist = [
     { name = "ebooklib", specifier = ">=0.18" },
     { name = "gewechat-client", specifier = ">=0.1.5" },
     { name = "html2text", specifier = ">=2024.2.26" },
-    { name = "langbot-plugin", specifier = "==0.3.6" },
+    { name = "langbot-plugin", specifier = "==0.3.8" },
     { name = "langchain", specifier = ">=0.2.0" },
     { name = "langchain-text-splitters", specifier = ">=0.0.1" },
     { name = "lark-oapi", specifier = ">=1.4.15" },
@@ -1993,7 +2009,7 @@ dev = [
 
 [[package]]
 name = "langbot-plugin"
-version = "0.3.6"
+version = "0.3.8"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "aiofiles" },
@@ -2011,9 +2027,9 @@ dependencies = [
     { name = "watchdog" },
     { name = "websockets" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/ff/f0/e5561bd1ebda0b9345ad6b98718b5f002bb3ca79b5ec294dc77cc10957b9/langbot_plugin-0.3.6.tar.gz", hash = "sha256:20db981e416a640f22246e54517abc2a095d8ccf5e69e06c2674fb8a443f5dbe", size = 179266, upload-time = "2026-03-30T15:58:58.523Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/b8/d8/7c8ac9516e35d69ead3e934b408e48541f5772eb88fbed19cd216af4b6c2/langbot_plugin-0.3.8.tar.gz", hash = "sha256:e8e420c3b2f167c9635e3e0af46fb452895be9d68ec05bf112ac5f221c3316f3", size = 179803, upload-time = "2026-04-10T11:05:42.791Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/a3/f5/ac424c2620e1be98a54a0b8ec0ed256a9c06cea7cd32a30732a1aea5fdc5/langbot_plugin-0.3.6-py3-none-any.whl", hash = "sha256:3238448436c41d50a0a0cf37438d845f0a1371159d440af3411a984e3d4e9eb7", size = 156752, upload-time = "2026-03-30T15:59:00.229Z" },
+    { url = "https://files.pythonhosted.org/packages/81/63/4a61b67d4886522647e0b60063da155279b943a6b2e6cd004e29aedf67d1/langbot_plugin-0.3.8-py3-none-any.whl", hash = "sha256:2246f343b4735cb4004cf44462ffb47531222c21efeef163a4acd758ebbec2cd", size = 157354, upload-time = "2026-04-10T11:05:41.525Z" },
 ]
 
 [[package]]
@@ -2409,6 +2425,18 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/6c/77/d7f491cbc05303ac6801651aabeb262d43f319288c1ea96c66b1d2692ff3/lxml-6.0.2-pp311-pypy311_pp73-win_amd64.whl", hash = "sha256:27220da5be049e936c3aca06f174e8827ca6445a4353a1995584311487fc4e3e", size = 3518768, upload-time = "2025-09-22T04:04:57.097Z" },
 ]
 
+[[package]]
+name = "mako"
+version = "1.3.10"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "markupsafe" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/9e/38/bd5b78a920a64d708fe6bc8e0a2c075e1389d53bef8413725c63ba041535/mako-1.3.10.tar.gz", hash = "sha256:99579a6f39583fa7e5630a28c3c1f440e4e97a414b80372649c0ce338da2ea28", size = 392474, upload-time = "2025-04-10T12:44:31.16Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/87/fb/99f81ac72ae23375f22b7afdb7642aba97c00a713c217124420147681a2f/mako-1.3.10-py3-none-any.whl", hash = "sha256:baef24a52fc4fc514a0887ac600f9f1cff3d82c61d4d700a1fa84d597b88db59", size = 78509, upload-time = "2025-04-10T12:50:53.297Z" },
+]
+
 [[package]]
 name = "markdown"
 version = "3.10.1"

web/.env.example

@@ -1 +1 @@
-NEXT_PUBLIC_API_BASE_URL=http://localhost:5300
+VITE_API_BASE_URL=http://localhost:5300

web/.gitignore

@@ -14,6 +14,7 @@
 /coverage
 
 # next.js
+/dist/
 /.next/
 /out/
 

web/components.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://ui.shadcn.com/schema.json",
   "style": "new-york",
-  "rsc": true,
+  "rsc": false,
   "tsx": true,
   "tailwind": {
     "config": "",

web/eslint.config.mjs

@@ -1,18 +1,27 @@
-import { dirname } from 'path';
-import { fileURLToPath } from 'url';
-import { FlatCompat } from '@eslint/eslintrc';
 import eslintPluginPrettierRecommended from 'eslint-plugin-prettier/recommended';
-
-const __filename = fileURLToPath(import.meta.url);
-const __dirname = dirname(__filename);
-
-const compat = new FlatCompat({
-  baseDirectory: __dirname,
-});
+import reactHooks from 'eslint-plugin-react-hooks';
+import tseslint from 'typescript-eslint';
 
 const eslintConfig = [
-  ...compat.extends('next/core-web-vitals', 'next/typescript'),
+  ...tseslint.configs.recommended,
+  {
+    files: ['**/*.{js,jsx,ts,tsx}'],
+    plugins: {
+      'react-hooks': reactHooks,
+    },
+    rules: {
+      ...reactHooks.configs.recommended.rules,
+      '@typescript-eslint/no-unused-vars': [
+        'warn',
+        { argsIgnorePattern: '^_', varsIgnorePattern: '^_' },
+      ],
+      '@typescript-eslint/no-explicit-any': 'off',
+    },
+  },
   eslintPluginPrettierRecommended,
+  {
+    ignores: ['dist/**', 'node_modules/**'],
+  },
 ];
 
 export default eslintConfig;

web/fix_router.sh

@@ -0,0 +1,2 @@
+sed -i 's/children={<HomePage \/>} />\n          <HomePage \/>\n        <\/HomeLayout>/g' src/router.tsx
+# well it's easier to recreate router.tsx

web/index.html

@@ -0,0 +1,16 @@
+<!doctype html>
+<html lang="zh">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>LangBot</title>
+    <meta
+      name="description"
+      content="Production-grade platform for building agentic IM bots"
+    />
+  </head>
+  <body>
+    <div id="root"></div>
+    <script type="module" src="/src/main.tsx"></script>
+  </body>
+</html>

web/migrate.sh

@@ -0,0 +1,29 @@
+#!/bin/bash
+cd /root/.openclaw/workspace/coding/projects/LangBot/web
+
+# Find and replace next/navigation
+find src -type f \( -name "*.ts" -o -name "*.tsx" \) -exec sed -i \
+  -e "s/import {.*useRouter.*} from 'next\/navigation'/import { useNavigate } from 'react-router-dom'/g" \
+  -e "s/import {.*usePathname.*} from 'next\/navigation'/import { useLocation } from 'react-router-dom'/g" \
+  -e "s/import {.*useSearchParams.*} from 'next\/navigation'/import { useSearchParams } from 'react-router-dom'/g" \
+  -e "s/const router = useRouter()/const navigate = useNavigate()/g" \
+  -e "s/router\.push(/navigate(/g" \
+  -e "s/router\.replace(/navigate(/g" \
+  -e "s/router\.back()/navigate(-1)/g" \
+  -e "s/router\.refresh()/navigate(0)/g" \
+  -e "s/const pathname = usePathname()/const location = useLocation();\n  const pathname = location.pathname;/g" \
+  -e "s/usePathname()/useLocation().pathname/g" \
+  {} +
+
+# Note: useSearchParams returns a tuple in react-router-dom. This might need manual fix depending on usage.
+
+# Replace next/link
+find src -type f \( -name "*.ts" -o -name "*.tsx" \) -exec sed -i \
+  -e "s/import Link from 'next\/link'/import { Link } from 'react-router-dom'/g" \
+  -e "s/<Link href=/<Link to=/g" \
+  {} +
+
+# Remove 'use client'
+find src -type f \( -name "*.ts" -o -name "*.tsx" \) -exec sed -i "s/'use client';//g" {} +
+find src -type f \( -name "*.ts" -o -name "*.tsx" \) -exec sed -i 's/"use client";//g' {} +
+

web/next.config.ts

@@ -1,8 +0,0 @@
-import type { NextConfig } from 'next';
-
-const nextConfig: NextConfig = {
-  /* config options here */
-  output: 'export',
-};
-
-export default nextConfig;

web/package.json

@@ -3,16 +3,15 @@
   "version": "0.1.0",
   "private": true,
   "scripts": {
-    "dev": "next dev --turbopack",
-    "build": "next build",
-    "start": "next start",
-    "lint": "eslint src",
-    "lint:fix": "eslint src --fix",
-    "lint-staged": "lint-staged"
+    "dev": "vite",
+    "build": "tsc && vite build",
+    "preview": "vite preview",
+    "lint": "eslint .",
+    "format": "prettier --write ."
   },
   "lint-staged": {
     "*.{js,jsx,ts,tsx}": [
-      "next lint --fix",
+      "eslint --fix",
       "prettier --write"
     ]
   },
@@ -46,17 +45,16 @@
     "@radix-ui/react-tooltip": "^1.2.7",
     "@tailwindcss/postcss": "^4.1.5",
     "@tanstack/react-table": "^8.21.3",
-    "axios": "^1.13.5",
+    "@vitejs/plugin-react": "^6.0.1",
+    "axios": "^1.15.0",
     "class-variance-authority": "^0.7.1",
     "clsx": "^2.1.1",
     "highlight.js": "^11.11.1",
     "i18next": "^25.1.2",
     "i18next-browser-languagedetector": "^8.1.0",
     "input-otp": "^1.4.2",
-    "lodash": "^4.17.23",
+    "lodash": "^4.18.0",
     "lucide-react": "^0.507.0",
-    "next": "~16.1.5",
-    "next-themes": "^0.4.6",
     "postcss": "^8.5.3",
     "qrcode": "^1.5.4",
     "react": "19.2.1",
@@ -65,6 +63,7 @@
     "react-i18next": "^15.5.1",
     "react-markdown": "^10.1.0",
     "react-photo-view": "^1.2.7",
+    "react-router-dom": "^7.14.0",
     "react-syntax-highlighter": "^16.1.0",
     "recharts": "2.15.4",
     "rehype-autolink-headings": "^7.1.0",
@@ -77,10 +76,10 @@
     "tailwind-merge": "^3.2.0",
     "tailwindcss": "^4.1.5",
     "uuidjs": "^5.1.0",
+    "vite": "^8.0.5",
     "zod": "^3.24.4"
   },
   "devDependencies": {
-    "@eslint/eslintrc": "^3",
     "@types/debug": "^4.1.12",
     "@types/estree": "^1.0.8",
     "@types/estree-jsx": "^1.0.5",
@@ -95,9 +94,10 @@
     "@types/react-syntax-highlighter": "^15.5.13",
     "@types/unist": "^3.0.3",
     "eslint": "^9",
-    "eslint-config-next": "15.2.4",
     "eslint-config-prettier": "^10.1.2",
     "eslint-plugin-prettier": "^5.2.6",
+    "eslint-plugin-react": "^7.37.5",
+    "eslint-plugin-react-hooks": "^5.2.0",
     "lint-staged": "^15.5.1",
     "prettier": "^3.5.3",
     "tw-animate-css": "^1.2.9",

web/scripts/check-i18n.mjs

@@ -0,0 +1,145 @@
+#!/usr/bin/env node
+/**
+ * Check that all i18n locale files have the same keys as en-US.ts (the reference).
+ * Reports missing keys (present in en-US but absent in the locale) and
+ * extra keys (present in the locale but absent in en-US).
+ * Exits with code 1 if any mismatch is found.
+ *
+ * Keys are extracted using a line-by-line parser that handles the known format
+ * of the locale files (no eval or dynamic code execution is used).
+ */
+
+import { readFileSync, readdirSync } from 'fs';
+import { resolve, dirname, join } from 'path';
+import { fileURLToPath } from 'url';
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const LOCALES_DIR = resolve(__dirname, '../src/i18n/locales');
+const REFERENCE = 'en-US.ts';
+
+/**
+ * Extract all dot-notation leaf keys from a TypeScript locale file.
+ *
+ * The expected file format is:
+ *   const <varName> = {
+ *     key: 'value',
+ *     nested: {
+ *       subKey: 'value',
+ *     },
+ *   };
+ *   export default <varName>;
+ *
+ * The parser tracks indentation depth to build dot-separated key paths and
+ * never executes the file content.
+ */
+function extractKeys(filePath) {
+  let src = readFileSync(filePath, 'utf8');
+
+  // Remove UTF-8 BOM if present
+  if (src.charCodeAt(0) === 0xfeff) {
+    src = src.slice(1);
+  }
+
+  const lines = src.split('\n');
+  const keys = [];
+  // Stack of { key, indent } pairs representing the current nesting path
+  const stack = [];
+
+  // Matches an object key at the start of a line (identifier or quoted string)
+  // Captures: [indent, keyName, hasOpenBrace]
+  const KEY_RE = /^(\s+)([\w]+)\s*:/;
+  const OPEN_BRACE_RE = /\{\s*$/;
+  const CLOSE_BRACE_RE = /^\s*\},?\s*$/;
+
+  for (const line of lines) {
+    if (CLOSE_BRACE_RE.test(line)) {
+      // Pop the stack when we encounter a closing brace line
+      const lineIndent = line.match(/^(\s*)/)[1].length;
+      while (stack.length > 0 && stack[stack.length - 1].indent >= lineIndent) {
+        stack.pop();
+      }
+      continue;
+    }
+
+    const m = line.match(KEY_RE);
+    if (!m) continue;
+
+    const indent = m[1].length;
+    const keyName = m[2];
+
+    // Pop stack entries that are at the same or deeper indent level
+    while (stack.length > 0 && stack[stack.length - 1].indent >= indent) {
+      stack.pop();
+    }
+
+    const prefix = stack.map((e) => e.key).join('.');
+    const fullKey = prefix ? `${prefix}.${keyName}` : keyName;
+
+    if (OPEN_BRACE_RE.test(line)) {
+      // This is a parent (nested object) key — push onto stack, don't record as leaf
+      stack.push({ key: keyName, indent });
+    } else {
+      // This is a leaf key
+      keys.push(fullKey);
+    }
+  }
+
+  return keys;
+}
+
+function main() {
+  const files = readdirSync(LOCALES_DIR).filter((f) => f.endsWith('.ts'));
+
+  if (!files.includes(REFERENCE)) {
+    console.error(`Reference file ${REFERENCE} not found in ${LOCALES_DIR}`);
+    process.exit(1);
+  }
+
+  const refKeys = new Set(extractKeys(join(LOCALES_DIR, REFERENCE)));
+  let hasError = false;
+
+  for (const file of files) {
+    if (file === REFERENCE) continue;
+
+    const locale = file.replace('.ts', '');
+    let localeKeys;
+    try {
+      localeKeys = new Set(extractKeys(join(LOCALES_DIR, file)));
+    } catch (e) {
+      console.error(`[${locale}] Failed to parse file: ${e.message}`);
+      hasError = true;
+      continue;
+    }
+
+    const missing = [...refKeys].filter((k) => !localeKeys.has(k));
+    const extra = [...localeKeys].filter((k) => !refKeys.has(k));
+
+    if (missing.length === 0 && extra.length === 0) {
+      console.log(`[${locale}] ✅ All keys match.`);
+    } else {
+      hasError = true;
+      console.log(`\n[${locale}] ❌ Key mismatch detected:`);
+      if (missing.length > 0) {
+        console.log(`  Missing keys (in en-US but not in ${locale}):`);
+        for (const k of missing) {
+          console.log(`    - ${k}`);
+        }
+      }
+      if (extra.length > 0) {
+        console.log(`  Extra keys (in ${locale} but not in en-US):`);
+        for (const k of extra) {
+          console.log(`    + ${k}`);
+        }
+      }
+    }
+  }
+
+  if (hasError) {
+    console.log('\n❌ i18n key check failed. Please fix the mismatches above.');
+    process.exit(1);
+  } else {
+    console.log('\n✅ All i18n locale files have matching keys.');
+  }
+}
+
+main();

web/src/app/auth/space/callback/page.tsx

@@ -1,7 +1,5 @@
-'use client';
-
 import { useEffect, useState, useCallback, Suspense } from 'react';
-import { useRouter, useSearchParams } from 'next/navigation';
+import { useNavigate, useSearchParams } from 'react-router-dom';
 import { httpClient } from '@/app/infra/http/HttpClient';
 import { toast } from 'sonner';
 import { useTranslation } from 'react-i18next';
@@ -23,8 +21,8 @@ import { LoadingSpinner } from '@/components/ui/loading-spinner';
 import langbotIcon from '@/app/assets/langbot-logo.webp';
 
 function SpaceOAuthCallbackContent() {
-  const router = useRouter();
-  const searchParams = useSearchParams();
+  const navigate = useNavigate();
+  const [searchParams] = useSearchParams();
   const { t } = useTranslation();
 
   const [status, setStatus] = useState<
@@ -51,7 +49,7 @@ function SpaceOAuthCallbackContent() {
         const wizardState = localStorage.getItem('langbot_wizard_state');
         const redirectTo = wizardState ? '/wizard' : '/home';
         setTimeout(() => {
-          router.push(redirectTo);
+          navigate(redirectTo);
         }, 1000);
       } catch (err) {
         setStatus('error');
@@ -64,7 +62,7 @@ function SpaceOAuthCallbackContent() {
         }
       }
     },
-    [router, t],
+    [navigate, t],
   );
 
   const [bindState, setBindState] = useState<string | null>(null);
@@ -81,7 +79,7 @@ function SpaceOAuthCallbackContent() {
         setStatus('success');
         toast.success(t('account.bindSpaceSuccess'));
         setTimeout(() => {
-          router.push('/home');
+          navigate('/home');
         }, 1000);
       } catch (err) {
         setStatus('error');
@@ -96,7 +94,7 @@ function SpaceOAuthCallbackContent() {
         setIsProcessing(false);
       }
     },
-    [router, t],
+    [navigate, t],
   );
 
   useEffect(() => {
@@ -146,7 +144,7 @@ function SpaceOAuthCallbackContent() {
   };
 
   const handleCancelBind = () => {
-    router.push('/home');
+    navigate('/home');
   };
 
   return (
@@ -154,7 +152,7 @@ function SpaceOAuthCallbackContent() {
       <Card className="w-[400px] shadow-lg dark:shadow-white/10">
         <CardHeader className="text-center">
           <img
-            src={langbotIcon.src}
+            src={langbotIcon}
             alt="LangBot"
             className="w-16 h-16 mb-4 mx-auto"
           />
@@ -217,7 +215,7 @@ function SpaceOAuthCallbackContent() {
             <>
               <AlertCircle className="h-12 w-12 text-red-500" />
               <Button
-                onClick={() => router.push(isBindMode ? '/home' : '/login')}
+                onClick={() => navigate(isBindMode ? '/home' : '/login')}
                 className="w-full mt-4"
               >
                 {isBindMode ? t('common.backToHome') : t('common.backToLogin')}