Google integration

Integrate with Google Gemini tools using LangChain JavaScript.

The @langchain/google package supports Gemini’s built-in tools, which provide capabilities like web search grounding, code execution, URL context retrieval, and more. These tools are passed as Gemini-native objects to ChatGoogle via bindTools() or the tools call option.

You cannot mix Gemini native tools (Google Search, Code Execution, etc.) with standard LangChain tools (Zod-based function tools) in the same request. See the ChatGoogle page for standard tool calling usage.

Google Search

The googleSearch tool grounds model responses with real-time Google Search results. This is useful for questions about current events or specific facts.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      googleSearch: {},
    },
  ]);

const res = await llm.invoke("Who won the latest World Series?");
console.log(res.text);

You can optionally filter search results to a specific time range:

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
  {
    googleSearch: {
      timeRangeFilter: {
        startTime: "2025-01-01T00:00:00Z",
        endTime: "2025-12-31T23:59:59Z",
      },
    },
  },
]);

The googleSearchRetrieval tool is maintained for backwards compatibility, but googleSearch is preferred.

For more information, see Google’s Grounding with Google Search documentation.

Code execution

The codeExecution tool allows Gemini to generate and run Python code to solve complex problems. The model writes the code, executes it, and returns the results.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      codeExecution: {},
    },
  ]);

const res = await llm.invoke("Calculate the 100th Fibonacci number.");
console.log(res.contentBlocks);

The response includes both the generated code and its execution result in the contentBlocks field:

for (const block of res.contentBlocks) {
  if (block.type === "tool_code") {
    console.log("Code:", block.toolCode);
  } else if (block.type === "tool_result") {
    console.log("Result:", block.toolResult);
  }
}

For more information, see Google’s Code Execution documentation.

URL context

The urlContext tool allows Gemini to fetch and use content from URLs to ground its responses.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      urlContext: {},
    },
  ]);

const res = await llm.invoke("Summarize this page: https://js.langchain.com/");
console.log(res.text);

For more information, see Google’s URL Context documentation.

Google Maps

The googleMaps tool grounds responses with geospatial context from Google Maps. This is useful for place-related queries.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      googleMaps: {},
    },
  ]);

const res = await llm.invoke("What are the best coffee shops near Times Square?");
console.log(res.text);

You can enable a widget context token for rendering a Google Maps widget:

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      googleMaps: {
        enableWidget: true,
    },
  },
]);

const res = await llm.invoke("Find Italian restaurants in downtown Chicago");

// Access the widget context token from grounding metadata
const groundingMetadata = res.response_metadata?.groundingMetadata;
console.log(groundingMetadata?.googleMapsWidgetContextToken);

For more information, see Google’s Google Maps grounding documentation.

File search

The fileSearch tool performs semantic retrieval from file search stores. Files must first be imported using the Gemini File API.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      fileSearch: {
      fileSearchStoreNames: ["fileSearchStores/my-store-123"],
    },
  },
]);

const res = await llm.invoke("What does the report say about Q4 revenue?");
console.log(res.text);

Configuration options:

• fileSearchStoreNames (required) — the names of the file search stores to retrieve from
• metadataFilter (optional) — metadata filter to apply to the retrieval
• topK (optional) — the number of semantic retrieval chunks to return

For more information, see Google’s File Search documentation.

Computer use

The computerUse tool enables Gemini to interact with a browser environment. The model can view screenshots and perform actions like clicking, typing, and scrolling.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      computerUse: {
      environment: "ENVIRONMENT_BROWSER",
    },
  },
]);

Configuration options:

• environment (required) — the environment being operated (e.g. "ENVIRONMENT_BROWSER")
• excludedPredefinedFunctions (optional) — predefined functions to exclude from the action space

For more information, see Google’s Computer Use documentation.

MCP servers

The mcpServers field allows Gemini to connect to remote MCP (Model Context Protocol) servers. Unlike other native tools, MCP servers are specified as an array on the tool object.

import { ChatGoogle } from "@langchain/google";

const llm = new ChatGoogle("gemini-2.5-flash")
  .bindTools([
    {
      mcpServers: [
      {
        name: "my-mcp-server",
        streamableHttpTransport: {
          url: "https://my-mcp-server.example.com/mcp",
        },
      },
    ],
  },
]);

const res = await llm.invoke("Use the tools from the MCP server to help me.");
console.log(res.text);

For more information, see Google’s MCP documentation.

Vertex AI Search data store

If you are using Vertex AI (platformType: "gcp"), you can ground responses using a Vertex AI Search data store.

import { ChatGoogle } from "@langchain/google";

const projectId = "YOUR_PROJECT_ID";
const datastoreId = "YOUR_DATASTORE_ID";

const llm = new ChatGoogle({
  model: "gemini-2.5-pro",
  platformType: "gcp",
}).bindTools([
  {
    retrieval: {
      vertexAiSearch: {
        datastore: `projects/${projectId}/locations/global/collections/default_collection/dataStores/${datastoreId}`,
      },
      disableAttribution: false,
    },
  },
]);

const res = await llm.invoke(
  "What is the score of Argentina vs Bolivia football game?"
);
console.log(res.text);

For more information, see Google’s Vertex AI Search grounding documentation.

---

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.

Edit this page on GitHub or file an issue.