Should a retrieved fact include its URL?

Keep the source URL and retrieval context with the result so the page can be checked again.

Firecrawl MCP Server

Official

Use Firecrawl MCP to retrieve website content and extract information for research through Firecrawl’s API-backed web tools.

Local serverstdioTypeScriptMIT★ 4.6k

What is the Firecrawl MCP Server?

Firecrawl MCP connects an assistant to Firecrawl’s website retrieval tools. It is useful when a research task needs readable page content or structured information from a defined set of URLs.

Specify the pages and fields you need

For a comparison, give the source URLs and requested fields, such as product name, published price and the date shown on the page. Keep the source link with each result. A successful scrape is a snapshot of retrieved content; it does not prove that a price, policy or availability will remain current.

Run the API-backed npm server

The documented local package is firecrawl-mcp. Register it with your client and set FIRECRAWL_API_KEY to a key from your Firecrawl account. Review the provider’s current tool list and account limits before a broad crawl. The npm server runs locally while its retrieval operations use Firecrawl’s service.

Setup and tool documentation. Explore related MCP servers.

How to install the Firecrawl MCP Server

{
  "mcpServers": {
    "firecrawl": {
      "command": "npx",
      "args": [
        "-y",
        "firecrawl-mcp"
      ],
      "env": {
        "FIRECRAWL_API_KEY": "YOUR_FIRECRAWL_API_KEY"
      }
    }
  }
}