browser-tools-mcp  by AgentDeskAI

Browser monitoring tool for AI coding assistants via MCP

Created 7 months ago
6,538 stars

Top 7.9% on SourcePulse

GitHubView on GitHub
1 Expert Loves This Project
Project Summary

This project provides a suite of browser auditing and monitoring tools for AI agents, enabling them to interact with and analyze web pages directly from IDEs like Cursor. It targets developers and AI engineers seeking to enhance their AI coding assistants with real-time browser context and analysis capabilities, offering automated checks for accessibility, performance, SEO, and best practices.

How It Works

The system comprises three components: a Chrome extension for capturing browser data (logs, DOM, screenshots), a Node.js server acting as middleware, and an MCP (Model Context Protocol) server integrated into the IDE. This architecture allows AI clients to send commands to the Node.js server via the MCP server, which then instructs the Chrome extension to perform actions like running audits or capturing specific data. Puppeteer and the Lighthouse library are leveraged for headless browser automation and detailed page analysis.

Quick Start & Requirements

Highlighted Details

  • Integrates Lighthouse for WCAG compliance, performance, SEO, and best practice audits.
  • Supports Next.js specific audits.
  • Offers "Debugger Mode" and "Audit Mode" for sequential tool execution.
  • Includes auto-paste functionality for screenshots into Cursor.
  • Resolves Windows connectivity issues and improves server/extension communication.

Maintenance & Community

  • Project roadmap available on GitHub.
  • Updates are regularly released (v1.2.0 noted).
  • Community engagement encouraged via issue tickets and X (@tedx_ai).

Licensing & Compatibility

  • License details are not explicitly stated in the README.
  • Compatible with any MCP-compatible client, primarily designed for Cursor.

Limitations & Caveats

The auto-paste feature requires the Agent input field in Cursor to be focused. The README mentions two servers that need to be installed and run separately, which could be a point of confusion for new users.

Health Check
Last Commit

5 months ago

Responsiveness

Inactive

Pull Requests (30d)
0
Issues (30d)
19
Star History
290 stars in the last 30 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), Luis Capelo Luis Capelo(Cofounder of Lightning AI), and
15 more.

stagehand by browserbase

0.5%
17k
AI browser automation framework for production
Created 1 year ago
Updated 1 day ago
Feedback? Help us improve.