Skip to main content

​
概述

LangChain 的 createAgent 底层运行在 LangGraph 的运行时上。 LangGraph 暴露了一个 Runtime 对象，包含以下信息：
  1. Context: 静态信息，例如用户 ID、数据库连接或其他代理调用所需的依赖项
  2. Store: 一个 BaseStore 实例，用于长期记忆
  3. Stream writer: 一个用于通过 "custom" 流模式流式传输信息的对象
您可以在工具内部和中间件内部访问运行时信息。

​
访问

当使用 createAgent 创建代理时，您可以指定一个 contextSchema 来定义存储在代理 Runtime 中的 context 的结构。 在调用代理时，通过 context 参数传递与本次运行相关的配置： 
import * as z from "zod";
import { createAgent } from "langchain";

const contextSchema = z.object({ 
  userName: z.string(), 
}); 

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

const result = await agent.invoke(
  { messages: [{ role: "user", content: "What's my name?" }] },
  { context: { userName: "John Smith" } } 
);

​
在工具内部

您可以在工具内部访问运行时信息，以便：
  • 访问上下文
  • 读取或写入长期记忆
  • 写入自定义流（例如，工具进度/更新）
使用 runtime 参数在工具内部访问 Runtime 对象。 
import * as z from "zod";
import { tool } from "langchain";
import { type Runtime } from "@langchain/langgraph"; 

const contextSchema = z.object({
  userName: z.string(),
});

const fetchUserEmailPreferences = tool(
  async (_, runtime: Runtime<z.infer<typeof contextSchema>>) => { 
    const userName = runtime.context?.userName; 
    if (!userName) {
      throw new Error("userName is required");
    }

    let preferences = "The user prefers you to write a brief and polite email.";
    if (runtime.store) { 
      const memory = await runtime.store?.get(["users"], userName); 
      if (memory) {
        preferences = memory.value.preferences;
      }
    }
    return preferences;
  },
  {
    name: "fetch_user_email_preferences",
    description: "Fetch the user's email preferences.",
    schema: z.object({}),
  }
);

​
在中间件内部

您可以在中间件中访问运行时信息，以创建动态提示、修改消息或根据用户上下文控制代理行为。 使用 runtime 参数在中间件内部访问 Runtime 对象。 
import * as z from "zod";
import { createAgent, createMiddleware, type AgentState, SystemMessage } from "langchain";
import { type Runtime } from "@langchain/langgraph"; 

const contextSchema = z.object({
  userName: z.string(),
});

// Dynamic prompt middleware
const dynamicPromptMiddleware = createMiddleware({
  name: "DynamicPrompt",
  beforeModel: (state: AgentState, runtime: Runtime<z.infer<typeof contextSchema>>) => {  
    const userName = runtime.context?.userName;  
    if (!userName) {
      throw new Error("userName is required");
    }

    const systemMsg = `You are a helpful assistant. Address the user as ${userName}.`;
    return {
      messages: [new SystemMessage(systemMsg), ...state.messages]
    };
  }
});

// Logging middleware
const loggingMiddleware = createMiddleware({
  name: "Logging",
  beforeModel: (state: AgentState, runtime: Runtime<z.infer<typeof contextSchema>>) => {  
    console.log(`Processing request for user: ${runtime.context?.userName}`);  
    return;
  },
  afterModel: (state: AgentState, runtime: Runtime<z.infer<typeof contextSchema>>) => {  
    console.log(`Completed request for user: ${runtime.context?.userName}`);  
    return;
  }
});

const agent = createAgent({
  model: "openai:gpt-4o",
  tools: [
    /* ... */
  ],
  middleware: [dynamicPromptMiddleware, loggingMiddleware],  
  contextSchema,
});

const result = await agent.invoke(
  { messages: [{ role: "user", content: "What's my name?" }] },
  { context: { userName: "John Smith" } }
);
Edit the source of this page on GitHub.
Connect these docs programmatically to Claude, VSCode, and more via MCP for real-time answers.