Firecrawl API SOP

Appearance

Firecrawl MCP + Cursor ​

Updated Feb 2026

Set up Firecrawl's MCP server in Cursor to give your AI assistant web scraping, searching, and crawling capabilities directly in the IDE.

Source: docs.firecrawl.dev/developer-guides/mcp-setup-guides/cursor

Prerequisites ​

Configuration ​

Add the Firecrawl MCP server to your Cursor settings. Open Settings > MCP and add the following configuration:

json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["-y", "firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-api-key"
      }
    }
  }
}

Configuration File Location

Cursor stores MCP configuration in your project's .cursor/mcp.json or globally in your Cursor settings. Check the Cursor docs for the latest path.

Windows Troubleshooting ​

If npx is not found on Windows, use the full path to npx.cmd:

json
{
  "mcpServers": {
    "firecrawl": {
      "command": "C:\\Program Files\\nodejs\\npx.cmd",
      "args": ["-y", "firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-api-key"
      }
    }
  }
}

Alternatively, find your npx.cmd path:

bash
where npx

Remote MCP (Alternative) ​

Instead of running the MCP server locally, you can use Firecrawl's remote MCP endpoint. This avoids needing Node.js installed:

json
{
  "mcpServers": {
    "firecrawl": {
      "url": "https://mcp.firecrawl.dev/YOUR_API_KEY_HERE/v2/mcp"
    }
  }
}

Replace YOUR_API_KEY_HERE with your Firecrawl API key.

Remote URL Contains Your Key

The remote MCP URL embeds your API key directly. Do not commit this to version control. Use environment variables or a .cursor/mcp.json file that is in your .gitignore.

Available Tools ​

Once configured, the following tools become available to Cursor's AI:

ToolDescription
firecrawl_scrapeScrape a single URL and return markdown
firecrawl_searchSearch the web and return results
firecrawl_crawlCrawl a website starting from a URL
firecrawl_mapDiscover all URLs on a website
firecrawl_extractExtract structured data from pages
firecrawl_batch_scrapeScrape multiple URLs in one request

Verify Setup ​

After adding the configuration:

  1. Restart Cursor or reload the window
  2. Open the AI chat panel
  3. Ask: "Scrape https://firecrawl.dev and tell me what it does"
  4. The AI should invoke the firecrawl_scrape tool and return a summary

Self-Hosted Option ​

If you run a self-hosted Firecrawl instance, add the FIRECRAWL_API_URL environment variable:

json
{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": ["-y", "firecrawl-mcp"],
      "env": {
        "FIRECRAWL_API_KEY": "fc-your-api-key",
        "FIRECRAWL_API_URL": "http://localhost:3002"
      }
    }
  }
}

Further Reading ​