> ## 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.

# 运行时

## 概述

LangChain 的 `createAgent` 底层运行在 LangGraph 的运行时之上。

LangGraph 暴露了一个 [`Runtime`](https://reference.langchain.com/javascript/langchain/index/Runtime) 对象，包含以下信息：

1. **上下文**：静态信息，如用户 ID、数据库连接或代理调用所需的其他依赖项
2. **存储**：一个 [BaseStore](https://reference.langchain.com/javascript/langchain-core/stores/BaseStore) 实例，用于[长期记忆](/oss/javascript/langchain/long-term-memory)
3. **流写入器**：一个对象，用于通过 `"custom"` 流模式流式传输信息

<Tip>
  运行时上下文是你将数据传递到代理中的方式。你可以将值（如数据库连接、用户会话或配置）附加到上下文，并在工具和中间件内部访问它们，而不是将内容存储在全局状态中。这保持了无状态、可测试和可重用的特性。
</Tip>

你可以在[工具内部](#inside-tools)和[中间件内部](#inside-middleware)访问运行时信息。

## 访问

使用 `createAgent` 创建代理时，你可以指定 `contextSchema` 来定义存储在代理 [`Runtime`](https://reference.langchain.com/javascript/langchain/index/Runtime) 中的 `context` 的结构。

调用代理时，传递包含运行相关配置的 `context` 参数：

```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import * as z from "zod";
import { createAgent } from "langchain";

const contextSchema = z.object({ // [!code highlight]
  userName: z.string(), // [!code highlight]
}); // [!code highlight]

const agent = createAgent({
  model: "gpt-4.1",
  tools: [
    /* ... */
  ],
  contextSchema, // [!code highlight]
});

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

### 在工具内部

你可以在工具内部访问运行时信息，以：

* 访问上下文
* 读取或写入长期记忆
* 写入[自定义流](/oss/javascript/langchain/streaming#custom-updates)（例如，工具进度/更新）

使用 `runtime` 参数在工具内部访问 [`Runtime`](https://reference.langchain.com/javascript/langchain/index/Runtime) 对象。

```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import * as z from "zod";
import { tool } from "langchain";
import { type ToolRuntime } from "@langchain/core/tools"; // [!code highlight]

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

const fetchUserEmailPreferences = tool(
  async (_, runtime: ToolRuntime<any, typeof contextSchema>) => { // [!code highlight]
    const userName = runtime.context?.userName; // [!code highlight]
    if (!userName) {
      throw new Error("userName is required");
    }

    let preferences = "用户希望你写一封简短而有礼貌的电子邮件。";
    if (runtime.store) { // [!code highlight]
      const memory = await runtime.store?.get(["users"], userName); // [!code highlight]
      if (memory) {
        preferences = memory.value.preferences;
      }
    }
    return preferences;
  },
  {
    name: "fetch_user_email_preferences",
    description: "获取用户的电子邮件偏好。",
    schema: z.object({}),
  }
);
```

### 在中间件内部

你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`](https://reference.langchain.com/javascript/langchain/index/Runtime) 对象。

```ts theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
import * as z from "zod";
import { createAgent, createMiddleware, SystemMessage } from "langchain";

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

// 动态提示中间件
const dynamicPromptMiddleware = createMiddleware({
  name: "DynamicPrompt",
  contextSchema,
  beforeModel: (state, runtime) => { // [!code highlight]
    const userName = runtime.context?.userName; // [!code highlight]
    if (!userName) {
      throw new Error("userName is required");
    }

    const systemMsg = `你是一个乐于助人的助手。请称呼用户为 ${userName}。`;
    return {
      messages: [new SystemMessage(systemMsg), ...state.messages],
    };
  },
});

// 日志中间件
const loggingMiddleware = createMiddleware({
  name: "Logging",
  contextSchema,
  beforeModel: (state, runtime) => {  // [!code highlight]
    console.log(`正在处理用户请求: ${runtime.context?.userName}`);  // [!code highlight]
    return;
  },
  afterModel: (state, runtime) => {  // [!code highlight]
    console.log(`已完成用户请求: ${runtime.context?.userName}`);  // [!code highlight]
    return;
  },
});

const agent = createAgent({
  model: "gpt-4.1",
  tools: [
    /* ... */
  ],
  middleware: [dynamicPromptMiddleware, loggingMiddleware],  // [!code highlight]
  contextSchema,
});

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

```

***

<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\runtime.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>