Skip to content
Get packs

Chrome MCP Server

A Chrome extension that lets AI assistants control and read your own open Chrome browser.hangwin/mcp-chrome

GitHub Download v1.0.0
Stars
12.5k
Forks
1.1k
Licence
MIT
Language
TypeScript
Last update
10 months ago
Type
App

What it is

Chrome MCP Server is a Chrome extension paired with a local Model Context Protocol (MCP) server. It lets AI assistants such as Claude read your open tabs, take screenshots, click and fill forms, and monitor network activity. Unlike Playwright-style automation, it works in your existing Chrome profile, so your logins and settings are already there.

  • Ask an AI assistant to list open tabs and summarize the content of a page
  • Take a screenshot of a page or a specific element through an AI client
  • Capture network requests or console output from a tab while debugging
  • Run a semantic search across the content of your open tabs

Great for

  • Developers who want an AI assistant to inspect or test pages in their everyday Chrome profile
  • People who already use logged-in sites and want an assistant to work within them
  • Users of MCP-compatible clients who want browser tools without launching a separate browser
  • Anyone experimenting with semantic search across open browser tabs

Not ideal for

  • Users who need fully isolated, repeatable test runs; the README contrasts this with Playwright's clean environments
  • Anyone uneasy about an AI acting inside a browser with active login sessions
  • People who need a stable, production-grade tool; the README says the project is in early stages
Step-by-step guide: how to use Chrome MCP ServerInstall, first run, examples, tips and FAQs

How to get it

  1. 1

    Install with a package manager

    You'll need: Node.js >= 20.0.0 · pnpm or npm · Chrome or Chromium browser

    Install the bridge globally with npm

    npm install -g mcp-chrome-bridge

    Install the bridge globally with pnpm (enable scripts first)

    pnpm install -g mcp-chrome-bridge

    Manually register the bridge if postinstall did not run

    mcp-chrome-bridge register
  2. 2

    Run it

    • Download the latest Chrome extension from the GitHub releases page
    • Open chrome://extensions/, enable Developer mode, and click Load unpacked on the extracted folder
    • Click the extension icon and click connect to see the MCP configuration
    • Add the Streamable HTTP config (url http://127.0.0.1:12306/mcp) to your MCP client and start using browser tools
  3. 3

    Get the most out of it

    • Prefer the Streamable HTTP connection, which the README marks as recommended over stdio
    • Use the semantic search tool (search_tabs_content) to find information across many open tabs
    • Use chrome_screenshot with element targeting or full-page options for visual checks
    • Read the full tool list in docs/TOOLS.md to see the 20+ available tools beyond those in the README excerpt
    • Review the VisualEditor documentation if you use Claude Code or Codex

Downloads

Source code (.zip)master branch · straight from GitHubRelease v1.0.01 files incl. chrome-mcp-server-1.0.0.zip

Strengths

  • Uses your existing Chrome browser, profile, and login state rather than a fresh browser
  • The README describes the server as fully local, which supports privacy
  • Works with any MCP-compatible client, with both Streamable HTTP and stdio connection options

Trade-offs

  • The README states the project is early-stage and under intensive development, so behavior may change
  • Setup involves a browser extension, a global npm package, and client configuration, which takes several steps
  • Because it acts inside your real logged-in browser, it carries security risk if an assistant misbehaves

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.