home / mcp / chrome mcp server

Chrome MCP Server

Chrome MCP Server is a Chrome extension-based Model Context Protocol (MCP) server that exposes your Chrome browser functionality to AI assistants like Claude, enabling complex browser automation, content analysis, and semantic search.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "hangwin-mcp-chrome": {
      "url": "http://127.0.0.1:12306/mcp"
    }
  }
}

Turn your Chrome browser into an intelligent, local assistant by running a Chrome MCP Server. This server exposes your Chrome browser functionality to AI assistants, enabling complex automation, content analysis, and semantic search directly in your existing browser environment while preserving privacy and login state.

How to use

You can connect an MCP client to control your Chrome browser through the Streamable HTTP method or via a local stdio workflow. The Streamable HTTP option is the recommended, straightforward path for ongoing interactions. The stdio approach runs a local process that your client can spawn and communicate with directly.

Connection options you can use from your MCP client include the HTTP streamable channel and a local stdio bridge. Use whichever matches your setup and preference. The HTTP method keeps the browser server accessible over a local HTTP endpoint, while the stdio method runs a local process that your client executes and talks to via standard I/O.

Below are explicit configuration examples you can copy into your MCP client setup. Adjust paths and URLs to fit your environment.

{
  "mcpServers": {
    "chrome-mcp-server": {
      "type": "streamableHttp",
      "url": "http://127.0.0.1:12306/mcp"
    }
  }
}

{
  "mcpServers": {
    "chrome-mcp-stdio": {
      "command": "npx",
      "args": [
        "node",
        "/Users/xxx/Library/pnpm/global/5/node_modules/mcp-chrome-bridge/dist/mcp/mcp-server-stdio.js"
      ]
    }
  }
}

Available tools

chrome_navigate

Navigate to URLs and control the browser viewport to move through pages programmatically.

chrome_switch_tab

Switch the active browser tab to a target tab by index or identifier.

chrome_close_tabs

Close specific tabs or entire browser windows as directed.

chrome_go_back_or_forward

Navigate backward or forward through the browser history.

chrome_inject_script

Inject a content script into a web page to modify behavior or extract data.

chrome_send_command_to_inject_script

Send commands from the MCP to an already injected script for dynamic control.

chrome_screenshot

Capture screenshots with options for elements, full-page capture, and custom dimensions.

chrome_network_capture_start/stop

Begin or end network capture using the webRequest API for monitoring traffic.

chrome_network_debugger_start/stop

Use the debugger API to inspect responses and bodies from network requests.

chrome_network_request

Send custom HTTP requests from the browser context.

search_tabs_content

AI-powered semantic search across content from open tabs.

chrome_get_web_content

Extract HTML or text content from a loaded page.

chrome_get_interactive_elements

Identify clickable or interactive elements on a page for automation.

chrome_console

Capture and retrieve console output from browser tabs.

chrome_history

Search browser history with support for time ranges.

chrome_bookmark_search

Find bookmarks by keywords across folders.

chrome_bookmark_add

Add new bookmarks and organize them into folders.

chrome_bookmark_delete

Delete bookmarks as needed.

chrome_click_element

Click elements using CSS selectors to perform page actions.

chrome_fill_or_select

Fill form fields and select options to interact with pages.

chrome_keyboard

Simulate keyboard input and shortcuts within the browser.