doc: add docs for knowledge base

2026-10-29 20:05:25 +08:00 · 2025-07-18 17:23:38 +08:00
parent 60fa2c3a94
commit 83cf595353
9 changed files with 82 additions and 5 deletions
@@ -128,7 +128,7 @@ export default defineConfig({
                ],
              },
              {
-                text: "配置对话模型",
+                text: "配置模型",
                link: "/zh/deploy/models/readme",
              },
              {
@@ -140,6 +140,10 @@ export default defineConfig({
                  { text: "n8n", link: "/zh/deploy/pipelines/n8n" },
                ],
              },
+              {
+                text: "使用知识库",
+                link: "/zh/deploy/knowledge/readme",
+              },
              {
                text: "系统环境设置",
                link: "/zh/deploy/settings",
@@ -354,6 +358,10 @@ export default defineConfig({
                  { text: "n8n", link: "/en/deploy/pipelines/n8n" },
                ],
              },
+              {
+                text: "Using Knowledge Base",
+                link: "/en/deploy/knowledge/readme",
+              },
              {
                text: "System Settings",
                link: "/en/deploy/settings",
@@ -0,0 +1,29 @@
+# Using Knowledge Base
+
+LangBot natively supports RAG (Retrieval-Augmented Generation), you can create a knowledge base and bind it to a pipeline, and the pipeline will answer questions based on the content in the knowledge base.
+
+:::info
+- You need to configure the embedding model before creating a knowledge base, please read [Configure Models](/en/deploy/models/readme).
+:::
+
+## Build Knowledge Base
+
+In the knowledge base page, click the `Create Knowledge Base` button, fill in the knowledge base name, select the configured embedding model, and click the `Create` button.
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/upload_docs.png" alt="create_kb" />
+
+## Use Knowledge Base
+
+Please go to the pipeline configuration, in the "AI" page, select `Local Agent` as the runner, and select the knowledge base you just created below.
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_local_agent.png" alt="use_kb" />
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_kb.png" alt="use_kb" />
+
+:::info
+Only when the runner is `Local Agent`, can you use the LangBot native knowledge base, if you want to use the knowledge base with other runners, please refer to the documentation of the corresponding product.
+:::
+
+Now you can use the knowledge base to chat in the `Chat Debug` or the robot bound to the pipeline:
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_kb_in_chat.png" alt="use_kb_in_chat" />
@@ -0,0 +1,30 @@
+# 使用知识库 (Knowledge Base)
+
+LangBot 原生支持 RAG（Retrieval-Augmented Generation，即检索增强生成），您可以创建知识库，并将其与流水线绑定，流水线将根据知识库中的内容回答问题。
+
+:::info
+- 知识库的创建和使用需要先配置嵌入模型，请先阅读[配置模型](/zh/deploy/models/readme)。
+:::
+
+## 构建知识库
+
+在知识库页面，点击`创建知识库`按钮，填写知识库名称，选择已配置好的嵌入模型，然后点击`创建`按钮。  
+之后请到知识库的“文档”页面，上传文档，LangBot 将会在后台开始解析和索引。
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/upload_docs.png" alt="create_kb" />
+
+## 使用知识库
+
+请到流水线配置中，“AI 能力”页，选择`内置 Agent`作为运行器，并在下方选择您刚刚创建的知识库。
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_local_agent.png" alt="use_kb" />
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_kb.png" alt="use_kb" />
+
+:::info
+仅当运行器为`内置 Agent`时，才可以使用 LangBot 原生知识库，若要在使用其他运行器时使用知识库，请参考所使用的运行器对应产品的文档。
+:::
+
+现在即可在`对话调试`或流水线所绑定的机器人上使用知识库进行对话：
+
+<img width="400px" src="/assets/image/zh/deploy/knowledge/use_kb_in_chat.png" alt="use_kb_in_chat" />
@@ -1,12 +1,14 @@
 # 配置模型 (Models)

-模型将被流水线用于处理消息，您配置的第一个模型将被设置为默认流水线的模型。
+## 对话模型 (LLM Model)
+
+对话模型将被流水线用于处理消息，您配置的第一个模型将被设置为默认流水线的模型。

 ![arch](/assets/image/zh/deploy/models/arch.png)

 可以添加多个模型，然后再流水线中选择具体使用什么模型

-![模型配置](/assets/image/zh/deploy/models/model_config.png)
+<img width="400px" src="/assets/image/zh/deploy/models/model_config.png" alt="model_config" />

 填入这四个参数`模型名称`、`模型供应商`、`请求 URL`、`API Key`，然后提交即可

@@ -26,6 +28,14 @@

 请求 URL 填入中转的 Base URL 即可，例如

-![模型配置](/assets/image/zh/deploy/models/other_provider.png)
+<img width="400px" src="/assets/image/zh/deploy/models/other_provider.png" alt="other_provider" />

-:::
+:::
+
+## 嵌入模型 (Embedding Model)
+
+嵌入模型将被用于计算消息的向量，若您需要使用知识库，请配置此模型。
+
+<img width="400px" src="/assets/image/zh/deploy/models/embedding_model.png" alt="embedding_model" />
+
+填入这四个参数`模型名称`、`模型供应商`、`请求 URL`、`API Key`，然后提交即可，之后请在知识库中配置使用此模型。