> ## Documentation Index
> Fetch the complete documentation index at: https://langchain-zh.cn/llms.txt
> Use this file to discover all available pages before exploring further.

# 自定义工作流

在**自定义工作流**架构中，你可以使用 [LangGraph](/oss/javascript/langgraph/overview) 定义自己的定制化执行流程。你可以完全控制图结构——包括顺序步骤、条件分支、循环和并行执行。

```mermaid theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
graph LR
    A([输入]) --> B{{条件判断}}
    B -->|路径_a| C[确定性步骤]
    B -->|路径_b| D((智能体步骤))
    C --> G([输出])
    D --> G([输出])

    classDef trigger fill:#DCFCE7,stroke:#16A34A,stroke-width:2px,color:#14532D
    classDef process fill:#DBEAFE,stroke:#2563EB,stroke-width:2px,color:#1E3A8A
    classDef decision fill:#FEF3C7,stroke:#F59E0B,stroke-width:2px,color:#78350F

    class A,G trigger
    class C,D process
    class B decision
```

## 主要特点

* 完全控制图结构
* 混合确定性逻辑与智能体行为
* 支持顺序步骤、条件分支、循环和并行执行
* 可将其他模式作为节点嵌入工作流中

## 使用场景

当标准模式（子智能体、技能等）不符合你的需求时，当你需要混合确定性逻辑与智能体行为时，或者你的用例需要复杂路由或多阶段处理时，请使用自定义工作流。

工作流中的每个节点可以是一个简单函数、一个LLM调用，或是一个完整的带有[工具](/oss/javascript/langchain/tools)的[智能体](/oss/javascript/langchain/agents)。你也可以在自定义工作流中组合其他架构——例如，将多智能体系统作为单个节点嵌入。

有关自定义工作流的完整示例，请参阅下面的教程。

<Card title="教程：构建具有路由功能的多源知识库" icon="book" href="/oss/javascript/langchain/multi-agent/router-knowledge-base" arrow cta="了解更多">
  [路由模式](/oss/javascript/langchain/multi-agent/router)是自定义工作流的一个示例。本教程将逐步指导你构建一个并行查询GitHub、Notion和Slack，然后综合结果的路由器。
</Card>

## 基础实现

核心思路是，你可以在任何LangGraph节点中直接调用LangChain智能体，从而将自定义工作流的灵活性与预构建智能体的便利性结合起来：

```typescript theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import { z } from "zod";
import { createAgent } from "langchain";
import { StateGraph, START, END, StateSchema, MessagesValue } from "@langchain/langgraph";

const agent = createAgent({ model: "openai:gpt-4o", tools: [...] });

const AgentState = new StateSchema({
  messages: MessagesValue,
  query: z.string(),
});

const agentNode: GraphNode<typeof AgentState> = (state) => {
  // 一个调用LangChain智能体的LangGraph节点
  const result = await agent.invoke({
    messages: [{ role: "user", content: state.query }]
  });
  return { answer: result.messages.at(-1)?.content };
}

// 构建一个简单的工作流
const workflow = new StateGraph(State)
  .addNode("agent", agentNode)
  .addEdge(START, "agent")
  .addEdge("agent", END)
  .compile();
```

## 示例：RAG管道

一个常见的用例是将[检索](/oss/javascript/langchain/retrieval)与智能体结合。这个示例构建了一个WNBA数据助手，它可以从知识库中检索信息，并能获取实时新闻。

<Accordion title="自定义RAG工作流">
  该工作流展示了三种类型的节点：

  * **模型节点**（重写）：使用[结构化输出](/oss/javascript/langchain/structured-output)重写用户查询以优化检索。
  * **确定性节点**（检索）：执行向量相似性搜索——不涉及LLM。
  * **智能体节点**（智能体）：基于检索到的上下文进行推理，并能通过工具获取额外信息。

  ```mermaid theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
  graph LR
      A([查询]) --> B{{重写}}
      B --> C[(检索)]
      C --> D((智能体))
      D --> E([响应])

      classDef trigger fill:#DCFCE7,stroke:#16A34A,stroke-width:2px,color:#14532D
      classDef process fill:#DBEAFE,stroke:#2563EB,stroke-width:2px,color:#1E3A8A

      class A,E trigger
      class B,C,D process
  ```

  <Tip>
    你可以使用LangGraph状态在工作流步骤之间传递信息。这允许工作流的每个部分读取和更新结构化字段，从而轻松地在节点之间共享数据和上下文。
  </Tip>

  ```typescript theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
  import { StateGraph, Annotation, START, END } from "@langchain/langgraph";
  import { createAgent, tool } from "langchain";
  import { ChatOpenAI, OpenAIEmbeddings } from "@langchain/openai";
  import { MemoryVectorStore } from "@langchain/classic/vectorstores/memory";
  import * as z from "zod";

  const State = Annotation.Root({
    question: Annotation<string>(),
    rewrittenQuery: Annotation<string>(),
    documents: Annotation<string[]>(),
    answer: Annotation<string>(),
  });

  // 包含阵容、比赛结果和球员数据的WNBA知识库
  const embeddings = new OpenAIEmbeddings();
  const vectorStore = await MemoryVectorStore.fromTexts(
    [
      // 阵容
      "New York Liberty 2024 roster: Breanna Stewart, Sabrina Ionescu, Jonquel Jones, Courtney Vandersloot.",
      "Las Vegas Aces 2024 roster: A'ja Wilson, Kelsey Plum, Jackie Young, Chelsea Gray.",
      "Indiana Fever 2024 roster: Caitlin Clark, Aliyah Boston, Kelsey Mitchell, NaLyssa Smith.",
      // 比赛结果
      "2024 WNBA Finals: New York Liberty defeated Minnesota Lynx 3-2 to win the championship.",
      "June 15, 2024: Indiana Fever 85, Chicago Sky 79. Caitlin Clark had 23 points and 8 assists.",
      "August 20, 2024: Las Vegas Aces 92, Phoenix Mercury 84. A'ja Wilson scored 35 points.",
      // 球员数据
      "A'ja Wilson 2024 season stats: 26.9 PPG, 11.9 RPG, 2.6 BPG. Won MVP award.",
      "Caitlin Clark 2024 rookie stats: 19.2 PPG, 8.4 APG, 5.7 RPG. Won Rookie of the Year.",
      "Breanna Stewart 2024 stats: 20.4 PPG, 8.5 RPG, 3.5 APG.",
    ],
    [{}, {}, {}, {}, {}, {}, {}, {}, {}],
    embeddings
  );
  const retriever = vectorStore.asRetriever({ k: 5 });

  const getLatestNews = tool(
    async ({ query }) => {
      // 你的新闻API在这里
      return "Latest: The WNBA announced expanded playoff format for 2025...";
    },
    {
      name: "get_latest_news",
      description: "获取最新的WNBA新闻和更新",
      schema: z.object({ query: z.string() }),
    }
  );

  const agent = createAgent({
    model: "openai:gpt-4.1",
    tools: [getLatestNews],
  });

  const model = new ChatOpenAI({ model: "gpt-4.1" });

  const RewrittenQuery = z.object({ query: z.string() });

  async function rewriteQuery(state: typeof State.State) {
    const systemPrompt = `重写此查询以检索相关的WNBA信息。
  知识库包含：球队阵容、带比分的比赛结果和球员统计数据（PPG、RPG、APG）。
  重点关注提到的具体球员姓名、球队名称或统计类别。`;
    const response = await model.withStructuredOutput(RewrittenQuery).invoke([
      { role: "system", content: systemPrompt },
      { role: "user", content: state.question },
    ]);
    return { rewrittenQuery: response.query };
  }

  async function retrieve(state: typeof State.State) {
    const docs = await retriever.invoke(state.rewrittenQuery);
    return { documents: docs.map((doc) => doc.pageContent) };
  }

  async function callAgent(state: typeof State.State) {
    const context = state.documents.join("\n\n");
    const prompt = `Context:\n${context}\n\nQuestion: ${state.question}`;
    const response = await agent.invoke({
      messages: [{ role: "user", content: prompt }],
    });
    return { answer: response.messages.at(-1)?.contentBlocks };
  }

  const workflow = new StateGraph(State)
    .addNode("rewrite", rewriteQuery)
    .addNode("retrieve", retrieve)
    .addNode("agent", callAgent)
    .addEdge(START, "rewrite")
    .addEdge("rewrite", "retrieve")
    .addEdge("retrieve", "agent")
    .addEdge("agent", END)
    .compile();

  const result = await workflow.invoke({
    question: "Who won the 2024 WNBA Championship?",
  });
  console.log(result.answer);
  ```
</Accordion>

***

<div className="source-links">
  <Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/i18n\zh-CN\oss\langchain\multi-agent\custom-workflow.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
  </Callout>

  <Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to Claude, VSCode, and more via MCP for real-time answers.
  </Callout>
</div>