CrawlMCP

A Model Context Protocol (MCP) server that fetches raw HTML content from Chrome pages via the Chrome DevTools Protocol. Designed to complement chrome-devtools-mcp by providing HTML extraction capabilities.

Features

• get_page_html: Fetch the complete raw HTML of any open Chrome page, or extract HTML of specific elements using CSS selectors
• list_chrome_pages: List all open pages in Chrome available for HTML extraction

Prerequisites

• Node.js 18+
• Chrome/Chromium running with remote debugging enabled

Start Chrome with Remote Debugging

# Windows
chrome.exe --remote-debugging-port=9222

# macOS
/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222

# Linux
google-chrome --remote-debugging-port=9222

Installation

npm install
npm run build

Usage

Configure in VS Code (Copilot)

Add to your VS Code settings.json:

{
  "mcp": {
    "servers": {
      "crawl-mcp": {
        "type": "stdio",
        "command": "npx",
        "args": [
          "-y",
          "crawl-mcp"
        ]
      }
    }
  }
}

Configure in Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "crawl-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "crawl-mcp"
      ]
    }
  }
}

Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | CDP_HOST | localhost | Chrome DevTools Protocol host | | CDP_PORT | 9222 | Chrome DevTools Protocol port |

Tools

get_page_html

Fetches raw HTML content from an open Chrome page.

Parameters:

• pageIndex (number, optional): Page index (0-based). Defaults to 0.
• selector (string, optional): CSS selector to get HTML of a specific element.

Examples:

Get full page HTML:

Use the get_page_html tool to fetch the HTML from the first Chrome tab

Get specific element:

Use get_page_html with selector "#main-content" to get only that section

list_chrome_pages

Lists all open pages accessible via Chrome DevTools Protocol.

Returns: Array of objects with index, title, and url.

Using with chrome-devtools-mcp

This MCP is designed to work alongside chrome-devtools-mcp. Use chrome-devtools-mcp for:

• Navigation and page interaction
• Taking screenshots and snapshots
• Console and network monitoring

Use CrawlMCP for:

• Extracting raw HTML source code
• Getting HTML of specific DOM elements

License

MIT