mcp-crawl4ai-ts

This MCP server provides a TypeScript implementation of a Crawl4AI-focused MCP server. It exposes a suite of tools for web crawling, content extraction, and browser automation, enabling clients to request content in structured formats and to automate common web data collection tasks. The server is designed to work with any MCP-compatible client by registering the crawl4ai MCP endpoint and providing a base URL to the Crawl4AI server. To start using it, run the MCP server via npx mcp-crawl4ai-ts (or install locally and run node dist/index.js), and configure your MCP client with the server name crawl4ai and the required base URL to Crawl4AI (CRAWL4AI_BASE_URL). The client can then invoke tools such as get_markdown, capture_screenshot, generate_pdf, execute_js, batch_crawl, smart_crawl, get_html, extract_links, crawl_recursive, parse_sitemap, and more, depending on your needs.

Prerequisites

• Node.js 18+ and npm
• A running Crawl4AI server (required for base URL integration)

Installation (via npx, recommended)

1. Ensure Node.js is installed (v18+).

2. No global installation is required; use npx to run the MCP server on demand:

   npx mcp-crawl4ai-ts

Configuration for clients

• In your MCP client configuration, reference the server as:

  { "mcpServers": { "crawl4ai": { "command": "npx", "args": ["mcp-crawl4ai-ts"], "env": { "CRAWL4AI_BASE_URL": "http://localhost:11235" } } } }

Optional: If you install locally and run the server directly, you can build and run from dist/index.js following the project’s build steps (e.g., npm install, npm run build) and then run with:

node dist/index.js

Prerequisites and environment variables are discussed in the Configuration section of the README.

Environment variables to consider:

• CRAWL4AI_BASE_URL: Required. Base URL of your running Crawl4AI server (e.g., http://localhost:11235).
• CRAWL4AI_API_KEY: Optional. If your server configuration requires authentication.
• SERVER_NAME: Optional. Custom name for this MCP server instance.
• SERVER_VERSION: Optional. Custom version for this MCP server instance.

Common issues:

• Ensure the Crawl4AI server is reachable at the CRAWL4AI_BASE_URL you provide; cross-origin or network restrictions can cause requests to fail.
• When using npx, the command will download the package on-demand; ensure your environment can access npm registries.
• If you install locally, remember to build TypeScript sources to dist before running.

Compatibility:

• Tested with Crawl4AI version 0.7.4 per the README notes. Verify compatibility with your Crawl4AI deployment before relying on all features.