ChromeDevTools / chrome-devtools-mcp

Gives AI coding agents full Chrome DevTools access via MCP, enabling browser debugging and automation through natural language

活跃维护 Apache 2.0 TypeScript Tracked
54.2k 5.9k 3 小时前
CINPM
mcp-server puppeteer chrome-devtools browser chrome debugging devtools mcp

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

chrome-devtools-mcp is an MCP server that gives AI coding agents (Claude, Cursor, Copilot, etc.) full access to Chrome DevTools for browser control, debugging, performance analysis, and automation. It combines Puppeteer-based automation with Chrome DevTools Protocol through the Model Context Protocol, enabling AI assistants to inspect network requests, take screenshots, analyze performance traces, and debug web applications.

为什么值得关注

With over 53K stars, this Google-backed project bridges the gap between AI coding agents and browser tooling, giving LLMs a standardized way to interact with live browsers. Its tight integration with Chrome DevTools provides rich debugging capabilities that pure text-based agents lack, making it a key enabler for AI-assisted web development workflows.

优势

  • Standardizes browser interaction for AI coding agents via MCP protocol
  • Leverages mature Chrome DevTools Protocol and Puppeteer for reliable automation
  • Strong development velocity with 98 contributors and 29 releases in 6 months
  • Excellent documentation with tool reference, design principles, and troubleshooting guides
  • Google-backed with official Chrome/Chrome for Testing support

局限性

  • Only officially supports Google Chrome, limiting cross-browser compatibility
  • No examples directory to help new users get started quickly
  • Exposes browser content to MCP clients, raising security/privacy concerns
  • No Docker support, complicating containerized deployments

使用场景

  • AI-assisted debugging of web applications via coding agents
  • Automated performance tracing and analysis during development
  • Network request inspection and screenshot capture for QA workflows
  • Teaching AI agents to navigate and test complex browser applications
目标用户: Developers using AI coding agents (Claude Code, Cursor, GitHub Copilot) who need browser interaction and debugging capabilities
学习曲线: 低
分析模型:LongCat-2.0 | 分析时间:12 天前