MCP_servers

This MCP server provides Firecrawl functionality, enabling web content extraction and scraping to be used within Cursor via the MCP (Model Context Protocol). Firecrawl acts as a modular tool that fetches and converts web pages into reusable content that can be fed into language models or other plugins. In Cursor, you configure the Firecrawl MCP server, authorize it with an API key, and then invoke its scraping capabilities from your prompts or workflows. The primary capability is to input a URL and receive structured output (e.g., extracted text or markdown) suitable for summarization, analysis, or citation in your conversations with the model.

To use Firecrawl through MCP, ensure the MCP server is running and Cursor is configured to communicate with the Firecrawl server. In Cursor, add a new MCP server with the name firecrawl, set the type to command, and point the command to firecrawl-mcp. Provide the required environment variable FIRECRAWL_API_KEY to enable authenticated access to the Firecrawl service. Once configured, you can issue requests from your prompts like: "Scrape this URL and return a markdown summary." The tool handles the web extraction behind the scenes and returns results that can be further processed by the model or saved for later use.

Prerequisites:

• Node.js (with npm) installed on your system
• Access to install global npm packages

Installation steps:

1. Install the Firecrawl MCP CLI globally: npm install -g firecrawl-mcp

2. (Optional) Verify installation: firecrawl-mcp --version

3. Configure Cursor to use the MCP server:

   • Ensure your Cursor configuration includes the mcp.json referencing the Firecrawl MCP as shown in the repository configuration.
   • Set the environment variable FIRECRAWL_API_KEY with your API key in Cursor or your environment where the MCP runs.

4. Start the MCP server (if needed by your setup): The command is firecrawl-mcp and it should run in the background or as part of your process manager depending on your deployment.

Prerequisites may vary slightly depending on your environment (local, Docker, or cloud). Ensure network access to the Firecrawl API and valid API credentials before use.

Tips and notes:

• API Key: You must provide FIRECRAWL_API_KEY to authenticate with the Firecrawl service.
• Cursor integration: Ensure mcp.json in your Cursor project correctly references the Firecrawl MCP and that Cursor is configured to load this file at startup.
• Dependency management: If you encounter permission issues when installing globally, consider using a Node version manager or installing with sudo (where appropriate for your OS).
• Rate limits: Be aware of any rate limits or usage quotas associated with your Firecrawl API key, and implement error handling for API errors in your prompts.
• Security: Never commit your API keys to version control. Use environment variables or secret management provided by your deployment platform.
• Logging: Enable verbose logs during setup to troubleshoot connection issues between Cursor and the Firecrawl MCP server.