Tool Calling

Tool calling in LangGraph.js connects model-generated tool requests to executable tools. ToolNode reads the tool calls in the latest AI message, runs the requested tools, and returns one ToolMessage per call; multiple calls can run in parallel.

new ToolNode(tools)

A Runnable node result containing ToolMessages, one for each executed tool call.

import { ToolNode } from "@langchain/langgraph/prebuilt";
import { tool } from "@langchain/core/tools";
import { z } from "zod";

const getWeather = tool(
  async ({ city }) => `Sunny in ${city}`,
  {
    name: "get_weather",
    description: "Get the weather for a city",
    schema: z.object({ city: z.string() }),
  },
);

const toolNode = new ToolNode([getWeather]);
const result = await toolNode.invoke({
  messages: [{
    role: "assistant",
    content: "",
    tool_calls: [{
      name: "get_weather",
      args: { city: "Paris" },
      id: "call-1",
      type: "tool_call",
    }],
  }],
});
console.log(result.messages[0].content);

ToolNode receives a message containing model-generated tool calls, executes the matching tools, and returns ToolMessages.