Repolode
Submit
eyalzh

eyalzh/browser-control-mcp

An MCP server and Firefox extension that lets AI assistants manage tabs, search history, and read or screenshot pages with user consent.

View on GitHub
Stars
336
Forks
80
Open issues
28
Last commit
2 months ago

What is it?

Browser Control MCP pairs a local MCP server with a Firefox extension so AI assistants can work inside the user's own browser. The assistant can open and close tabs, list open tabs, create named and colored tab groups, reorder tabs, and read or search browsing history.

For page content, the assistant can read a webpage's text and links, find and highlight text in a tab, and capture a screenshot of a tab's visible area. Reading page content requires the user's consent per domain, and screenshots require the user to authorize each tab by clicking the extension's toolbar button, enforced through Firefox's activeTab permission.

The design deliberately avoids page modification, arbitrary scripting, and remote data collection. The MCP server and extension communicate over a local connection secured by a shared secret, and the extension keeps an audit log of tool calls. The project is marked experimental.

Key features

  • 01Open, close, and list browser tabs
  • 02Create tab groups with a name and color
  • 03Reorder open tabs
  • 04Read and search browsing history
  • 05Read page text and links with per-domain consent
  • 06Capture screenshots of a tab after per-tab authorization

Use cases

  • Closing or grouping tabs by topic with an AI assistant
  • Searching past browsing history for a page you remember
  • Having an assistant research across several search results and summarize them
  • Reviewing a webpage's content with the assistant without granting it page-editing access

Quick start

  1. 1

    Install the Firefox extension from addons.mozilla.org

  2. 2

    Clone the repository and install dependencies

    npm install
  3. 3

    Build the MCP server and browser extension

    npm run build
  4. 4

    Add the MCP server configuration with the extension secret to your MCP client config

Things to know

  • The project is marked experimental and the README says to use it at your own risk.
  • Firefox only; it does not work with other browsers.
  • Setup requires copying a secret key from the extension page into the MCP server configuration.

Topics

#mcp#browser-extension#firefox#ai-agent#browser-automation#tabs

FAQ

What is browser-control-mcp?
An MCP server and Firefox extension that lets AI assistants manage tabs, search history, and read or screenshot pages with user consent. Browser Control MCP pairs a local MCP server with a Firefox extension so AI assistants can work inside the user's own browser. The assistant can open and close tabs, list open tabs, create named and colored tab groups, reorder tabs, and read or search browsing history.
Is browser-control-mcp free to use? What license does it have?
browser-control-mcp is open source under the MIT license. Check the license terms before using it in commercial products.
Is browser-control-mcp still maintained?
Updated within 6 months. The last commit to browser-control-mcp was on Aug 23, 2026.
What can I use browser-control-mcp for?
Closing or grouping tabs by topic with an AI assistant Searching past browsing history for a page you remember Having an assistant research across several search results and summarize them Reviewing a webpage's content with the assistant without granting it page-editing access
How do I install browser-control-mcp?
Install the Firefox extension from addons.mozilla.org Clone the repository and install dependencies: npm install Build the MCP server and browser extension: npm run build Add the MCP server configuration with the extension secret to your MCP client config
Where can I get browser-control-mcp?
The source code and installation instructions are on GitHub: https://github.com/eyalzh/browser-control-mcp

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects