Skip to main content
LangGraph 提供了两种不同的 API 来构建 Agent 工作流：Graph API 和 Functional API。这两种 API 共享相同的底层运行时，并且可以在同一个应用程序中一起使用，但它们是为不同的用例和开发偏好而设计的。 本指南将帮助您根据您的具体需求了解何时使用每种 API。

​
快速决策指南

当您需要以下内容时，请使用 Graph API：
  • 复杂的工作流可视化 用于调试和文档
  • 显式状态管理，在多个节点之间共享数据
  • 条件分支，具有多个决策点
  • 并行执行路径，稍后需要合并
  • 团队协作，其中可视化表示有助于理解
当您想要以下内容时，请使用 Functional API：
  • 对现有过程代码进行 最小的代码更改
  • 标准控制流（if/else、循环、函数调用）
  • 函数作用域状态，无需显式状态管理
  • 快速原型设计，样板代码更少
  • 线性工作流，具有简单的分支逻辑

​
详细比较

​
何时使用 Graph API

Graph API 使用声明式方法，您定义节点、边和共享状态以创建可视化的图结构。 1. 复杂的决策树和分支逻辑 当您的工作流具有依赖于各种条件的多个决策点时，Graph API 使这些分支变得显式且易于可视化。 
import * as z from "zod";
import {
  StateGraph,
  StateSchema,
  MessagesValue,
  START,
  END,
  type GraphNode,
  type ConditionalEdgeRouter,
} from "@langchain/langgraph";

// Graph API: Clear visualization of decision paths
const AgentState = new StateSchema({
  messages: MessagesValue,
  currentTool: z.string(),
  retryCount: z.number().default(0),
});

const shouldContinue: ConditionalEdgeRouter<typeof AgentState> = (state) => {
  if (state.retryCount > 3) {
    return END;
  } else if (state.currentTool === "search") {
    return "processSearch";
  } else {
    return "callLlm";
  }
};

const workflow = new StateGraph(AgentState)
  .addNode("callLlm", callLlmNode)
  .addNode("processSearch", searchNode)
  .addConditionalEdges("callLlm", shouldContinue);
2. 跨多个组件的状态管理 当您需要在工作流的不同部分之间共享和协调状态时，Graph API 的显式状态管理非常有益。 
import * as z from "zod";
import { StateSchema, type GraphNode } from "@langchain/langgraph";

// Multiple nodes can access and modify shared state
const WorkflowState = new StateSchema({
  userInput: z.string(),
  searchResults: z.array(z.string()).default([]),
  generatedResponse: z.string().optional(),
  validationStatus: z.string().optional(),
});

const searchNode: GraphNode<typeof WorkflowState> = async (state) => {
  // Access shared state
  const results = await search(state.userInput);
  return { searchResults: results };
};

const validationNode: GraphNode<typeof WorkflowState> = async (state) => {
  // Access results from previous node
  const isValid = await validate(state.generatedResponse);
  return { validationStatus: isValid ? "valid" : "invalid" };
};
3. 具有同步的并行处理 当您需要并行运行多个操作然后合并它们的结果时，Graph API 可以自然地处理这种情况。 
import { START } from "@langchain/langgraph";

// Parallel processing of multiple data sources
workflow
  .addNode("fetchNews", fetchNews)
  .addNode("fetchWeather", fetchWeather)
  .addNode("fetchStocks", fetchStocks)
  .addNode("combineData", combineAllData)
  // All fetch operations run in parallel
  .addEdge(START, "fetchNews")
  .addEdge(START, "fetchWeather")
  .addEdge(START, "fetchStocks")
  // Combine waits for all parallel operations to complete
  .addEdge("fetchNews", "combineData")
  .addEdge("fetchWeather", "combineData")
  .addEdge("fetchStocks", "combineData");
4. 团队开发和文档 Graph API 的可视化特性使团队更容易理解、记录和维护复杂的工作流。 
// Clear separation of concerns - each team member can work on different nodes
workflow
  .addNode("dataIngestion", dataTeamFunction)
  .addNode("mlProcessing", mlTeamFunction)
  .addNode("businessLogic", productTeamFunction)
  .addNode("outputFormatting", frontendTeamFunction);

​
何时使用 Functional API

Functional API 使用命令式方法，将 LangGraph 功能集成到标准过程代码中。 1. 现有的过程代码 当您拥有使用标准控制流的现有代码，并希望以最小的重构添加 LangGraph 功能时。 
import { task, entrypoint } from "@langchain/langgraph";

// Functional API: Minimal changes to existing code
const processUserInput = task(
  "processUserInput",
  async (userInput: string) => {
    // Existing function with minimal changes
    return { processed: userInput.toLowerCase().trim() };
  }
);

const workflow = entrypoint(
  { checkpointer },
  async (userInput: string) => {
    // Standard control flow
    const processed = await processUserInput(userInput);

    let response: string;
    if (processed.processed.includes("urgent")) {
      response = await handleUrgentRequest(processed);
    } else {
      response = await handleNormalRequest(processed);
    }

    return response;
  }
);
2. 逻辑简单的线性工作流 当您的工作流主要是顺序的，并具有简单的条件逻辑时。 
import { entrypoint, interrupt } from "@langchain/langgraph";

const essayWorkflow = entrypoint(
  { checkpointer },
  async (topic: string) => {
    // Linear flow with simple branching
    let outline = await createOutline(topic);

    if (outline.points.length < 3) {
      outline = await expandOutline(outline);
    }

    const draft = await writeDraft(outline);

    // Human review checkpoint
    const feedback = interrupt({ draft, action: "Please review" });

    let finalEssay: string;
    if (feedback === "approve") {
      finalEssay = draft;
    } else {
      finalEssay = await reviseEssay(draft, feedback);
    }

    return { essay: finalEssay };
  }
);
3. 快速原型设计 当您想快速测试想法而无需定义状态模式和图结构的开销时。 
import { entrypoint } from "@langchain/langgraph";

const quickPrototype = entrypoint(
  { checkpointer },
  async (data: Record<string, unknown>) => {
    // Fast iteration - no state schema needed
    const step1Result = await processStep1(data);
    const step2Result = await processStep2(step1Result);

    return { finalResult: step2Result };
  }
);
4. 函数作用域的状态管理 当您的状态自然地限定在单个函数范围内，并且不需要广泛共享时。 
import { task, entrypoint } from "@langchain/langgraph";

const analyzeDocument = task("analyzeDocument", async (document: string) => {
  // Local state management within function
  const sections = extractSections(document);
  const summaries = await Promise.all(sections.map(summarize));
  const keyPoints = extractKeyPoints(summaries);

  return {
    sections: sections.length,
    summaries,
    keyPoints,
  };
});

const documentProcessor = entrypoint(
  { checkpointer },
  async (document: string) => {
    const analysis = await analyzeDocument(document);
    // State is passed between functions as needed
    return await generateReport(analysis);
  }
);

​
结合使用两种 API

您可以在同一个应用程序中一起使用这两种 API。当系统的不同部分有不同的要求时，这很有用。 
import * as z from "zod";
import {
  StateGraph,
  StateSchema,
  entrypoint,
  type GraphNode,
} from "@langchain/langgraph";

// Define state for complex multi-agent coordination
const CoordinationState = new StateSchema({
  rawData: z.record(z.string(), z.unknown()),
  processedData: z.record(z.string(), z.unknown()).optional(),
});

// Simple data processing using Functional API
const dataProcessor = entrypoint({}, async (rawData: Record<string, unknown>) => {
  const cleaned = await cleanData(rawData);
  const transformed = await transformData(cleaned);
  return transformed;
});

// Use the functional API result in the graph
const orchestratorNode: GraphNode<typeof CoordinationState> = async (state) => {
  const processedData = await dataProcessor.invoke(state.rawData);
  return { processedData };
};

// Complex multi-agent coordination using Graph API
const coordinationGraph = new StateGraph(CoordinationState)
  .addNode("orchestrator", orchestratorNode)
  .addNode("agentA", agentANode)
  .addNode("agentB", agentBNode);

​
API 之间的迁移

​
从 Functional API 迁移到 Graph API

当您的函数式工作流变得复杂时，您可以迁移到 Graph API： 
import * as z from "zod";
import { entrypoint } from "@langchain/langgraph";

// Before: Functional API
const complexWorkflow = entrypoint(
  { checkpointer },
  async (inputData: Record<string, unknown>) => {
    const step1 = await processStep1(inputData);

    let result: unknown;
    if (step1.needsAnalysis) {
      const analysis = await analyzeData(step1);
      if (analysis.confidence > 0.8) {
        result = await highConfidencePath(analysis);
      } else {
        result = await lowConfidencePath(analysis);
      }
    } else {
      result = await simplePath(step1);
    }

    return result;
  }
);

// After: Graph API
import {
  StateGraph,
  StateSchema,
  type GraphNode,
  type ConditionalEdgeRouter,
} from "@langchain/langgraph";

const WorkflowState = new StateSchema({
  inputData: z.record(z.string(), z.unknown()),
  step1Result: z.record(z.string(), z.unknown()).optional(),
  analysis: z.record(z.string(), z.unknown()).optional(),
  finalResult: z.unknown().optional(),
});

const shouldAnalyze: ConditionalEdgeRouter<typeof WorkflowState> = (state) => {
  return state.step1Result?.needsAnalysis ? "analyze" : "simplePath";
};

const confidenceCheck: ConditionalEdgeRouter<typeof WorkflowState> = (state) => {
  return (state.analysis?.confidence as number) > 0.8
    ? "highConfidence"
    : "lowConfidence";
};

const workflow = new StateGraph(WorkflowState)
  .addNode("step1", processStep1Node)
  .addConditionalEdges("step1", shouldAnalyze)
  .addNode("analyze", analyzeDataNode)
  .addConditionalEdges("analyze", confidenceCheck);
// ... add remaining nodes and edges

​
从 Graph API 迁移到 Functional API

当您的图对于简单的线性流程变得过于复杂时： 
import { z } from "zod/v4";
import { StateGraph, StateSchema, entrypoint } from "@langchain/langgraph";

// Before: Over-engineered Graph API
const SimpleState = new StateSchema({
  input: z.string(),
  step1: z.string().optional(),
  step2: z.string().optional(),
  result: z.string().optional(),
});

// After: Simplified Functional API
const simpleWorkflow = entrypoint(
  { checkpointer },
  async (inputData: string) => {
    const step1 = await processStep1(inputData);
    const step2 = await processStep2(step1);
    return await finalizeResult(step2);
  }
);

​
总结

当您需要对工作流结构、复杂分支、并行处理或团队协作优势进行显式控制时，请选择 Graph API。 当您希望以最小的更改将 LangGraph 功能添加到现有代码、具有简单的线性工作流或需要快速原型设计功能时，请选择 Functional API。 这两种 API 都提供相同的核心 LangGraph 功能（持久性、流式传输、人机交互、记忆），但将它们封装在不同的范式中，以适应不同的开发风格和用例。
在 GitHub 上编辑此页面 或 提交 issue。
将本文档连接 到 Claude、VSCode 等，通过 MCP 获取实时解答。