puppeteer-real-browser

This MCP server lets AI assistants interact with a real web browser through Puppeteer Real Browser, enabling actions like navigating to web pages, clicking elements, filling forms, extracting content, and solving certain automation tasks with anti-detection features. The server exposes a set of tools designed for browser automation and data extraction, making it possible for copilots to perform complex tasks in a real browser while maintaining stealth and stability. After starting or configuring the server, you can instruct your AI assistant to open websites, perform searches, fill forms, scrape content, handle logins, and automate multi-step workflows across sites. The included capabilities cover navigation, interaction, data extraction, captcha handling, scrolling, proxy support, and robust error recovery to ensure reliable operation in diverse environments.

Prerequisites:

• Node.js version 18 or higher
• npm or yarn
• Google Chrome or Chromium installed
• Basic understanding of JavaScript/TypeScript (for development)

For end users (no permanent installation required):

1. Ensure Node.js is installed (version 18+): node --version
2. Configure Claude Desktop to use the MCP server via npx (no local install needed). The README provides the exact JSON configuration used by Claude Desktop: { "mcpServers": { "puppeteer-real-browser": { "command": "npx", "args": ["puppeteer-real-browser-mcp-server@latest"] } } }

For developers who want to run locally or contribute:

1. Install the package globally (optional, for command-line usage):

   npm install -g puppeteer-real-browser-mcp-server@latest
   

2. Run the server directly (if installed globally):

   puppeteer-real-browser-mcp-server
   

3. Or clone the repo and run locally:

   git clone https://github.com/withLinda/puppeteer-real-browser-mcp-server.git
   cd puppeteer-real-browser-mcp-server
   npm install
   npm run build
   npm run dev
   

Tips and common configurations:

• If using Claude Desktop, the recommended approach is to rely on npx to fetch the latest release, ensuring bug fixes and improvements are applied automatically.
• You can customize the browser behavior with puppeteer-real-browser options, including headless mode (headless: true) and CHROME_PATH for specific Chrome installations.
• For Windows users, the readme lists various Chrome installation paths and can leverage CHROME_PATH to specify the binary location.
• When troubleshooting, check that Node.js is up to date (18+), that Chrome/Chromium is installed, and that the MCP server configuration points to the latest package version.
• If you encounter network-related limitations or bot-detection measures, review the proxy settings and ensure appropriate network access for the browser instance.