MCP.so
ログイン

BrowserTools MCP

@AgentDeskAI

BrowserTools MCP について

Monitor browser logs directly from Cursor and other MCP compatible IDEs.

基本情報

カテゴリ

ブラウザ自動化

ライセンス

MIT

ランタイム

node

トランスポート

stdio

公開者

AgentDeskAI

設定

標準の設定はありません

このサーバーの README には解析可能な MCP 設定ブロックが含まれていません。インストール手順はリポジトリをご確認ください。

リポジトリ

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is BrowserTools MCP?

BrowserTools MCP is a browser monitoring and interaction tool that enables AI-powered applications via Anthropic’s Model Context Protocol (MCP) to capture and analyze browser data through a Chrome extension. It consists of three components: a Chrome extension, a Node.js middleware server, and an MCP server. It is designed for developers using MCP-compatible clients (e.g., Cursor, Cline, Zed, Claude Desktop) who want to give their AI tools real-time access to browser console logs, network traffic, screenshots, DOM elements, and automated audits.

How to use BrowserTools MCP?

Install the Chrome extension (v1.2.0), then run npx @agentdeskai/browser-tools-mcp@latest inside your IDE and npx @agentdeskai/browser-tools-server@latest in a separate terminal. Open Chrome DevTools and locate the BrowserToolsMCP panel. If issues arise, close Chrome completely, restart the local server, and ensure only one DevTools instance is open.

Key features of BrowserTools MCP

  • Captures console logs, network activity, and DOM elements in real time
  • Takes full-page screenshots via WebSocket commands
  • Runs accessibility (WCAG), performance, SEO, and best practices audits
  • NextJS-specific audit and combined audit/debugger modes
  • Stores all data locally; never sent to third parties
  • Auto-discovers hosts/ports, supports auto-reconnect and graceful shutdown

Use cases of BrowserTools MCP

  • Let an AI coding agent debug browser console errors without switching contexts
  • Automate WCAG compliance checks during development in the IDE
  • Quickly identify performance bottlenecks (e.g., render‑blocking resources, oversized images)
  • Improve on‑page SEO by triggering Lighthouse‑based audits from an AI chat

FAQ from BrowserTools MCP

Why does the README say the project is no longer active?

The README states “THIS PROJECT IS NO LONGER ACTIVE PLEASE USE A DIFFERENT SOLUTION FOR THIS.” It is deprecated; the information below reflects the state at the time of archive.

What are the runtime dependencies and installation requirements?

You need the Chrome BrowserTools extension (v1.2.0 from GitHub releases) and Node.js. Two packages must be installed: @agentdeskai/browser-tools-mcp (MCP server for your IDE) and @agentdeskai/browser-tools-server (Node middleware server).

Where does browser data live?

All logs and captured data are stored locally on your machine. The servers never send data to any third‑party service or API.

What transport and authentication does the server use?

The server uses Websocket connections between the Chrome extension and Node server, and standard MCP transport for communication with the MCP client. The README does not detail authentication; no API keys are mentioned.

Which MCP clients are compatible?

The system works with any MCP‑compatible client, including Cursor, Cline, Zed, and Claude Desktop. Cursor is explicitly mentioned as a primary target.

コメント

「ブラウザ自動化」の他のコンテンツ