Repolode
Submit
firecrawl

firecrawl/firecrawl-mcp-server

Official MCP server that gives AI agents and MCP clients web search, scraping, crawling, browser interaction and Markdown or structured-data extraction via Firecrawl.

View on GitHubWebsite
Stars
7,642
Forks
934
Open issues
175
Last commit
4 hours ago

What is it?

Firecrawl MCP Server is a Model Context Protocol server that exposes Firecrawl's web tools to any MCP-compatible client such as Claude Desktop, Cursor, Windsurf or VS Code. Agents can search the web, scrape a URL into clean Markdown or JSON matching a schema, crawl or map a site, parse PDFs and documents, and drive pages in a real browser with clicks, typing and navigation.

The server also covers higher-level workflows: multi-source Agent jobs that return structured research, recurring Monitor checks with change alerts, developer, government and scientific-paper search indexes, and credit usage reporting. It runs over stdio via npx or Docker, or as a Streamable HTTP server, and can point at a self-hosted Firecrawl API through FIRECRAWL_API_URL.

Key features

  • 01Search the web, optionally returning full page content for each result
  • 02Scrape URLs into Markdown or structured JSON using a supplied schema
  • 03Crawl a site or map its URLs with limits on depth and paths
  • 04Interact with dynamic pages in a browser, including clicks, forms and navigation
  • 05Parse PDFs and other documents into Markdown or JSON
  • 06Runs over stdio or Streamable HTTP, with support for self-hosted Firecrawl

Use cases

  • Giving a Claude or Cursor assistant live access to documentation or product pages
  • Extracting product or listing data from known URLs into a fixed JSON schema
  • Discovering and collecting all pages under a documentation site for indexing
  • Monitoring a web page for content changes on a schedule

Quick start

  1. 1

    Install the server globally with npm

    npm install -g firecrawl-mcp
  2. 2

    Run the server with your Firecrawl API key

    env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp
  3. 3

    Or connect to the hosted server with no setup (keyless free tier)

Things to know

  • Most tools, including crawl, map and agent, need a Firecrawl API key; the keyless hosted tier offers only a few tools.
  • Each browser interaction call runs one turn to completion, so you cannot drive a session step by step from the client.

Topics

#mcp#web-scraping#web-crawler#llm-tools#content-extraction#search-api#ai-agents#batch-processing#claude#data-collection

FAQ

What is firecrawl-mcp-server?
Official MCP server that gives AI agents and MCP clients web search, scraping, crawling, browser interaction and Markdown or structured-data extraction via Firecrawl. Firecrawl MCP Server is a Model Context Protocol server that exposes Firecrawl's web tools to any MCP-compatible client such as Claude Desktop, Cursor, Windsurf or VS Code. Agents can search the web, scrape a URL into clean Markdown or JSON matching a schema, crawl or map a site, parse PDFs and documents, and drive pages in a real browser with clicks, typing and navigation.
Is firecrawl-mcp-server free to use? What license does it have?
firecrawl-mcp-server is open source under the MIT license. Check the license terms before using it in commercial products.
Is firecrawl-mcp-server still maintained?
Actively maintained. The last commit to firecrawl-mcp-server was on Oct 20, 2026.
What can I use firecrawl-mcp-server for?
Giving a Claude or Cursor assistant live access to documentation or product pages Extracting product or listing data from known URLs into a fixed JSON schema Discovering and collecting all pages under a documentation site for indexing Monitoring a web page for content changes on a schedule
How do I install firecrawl-mcp-server?
Install the server globally with npm: npm install -g firecrawl-mcp Run the server with your Firecrawl API key: env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp Or connect to the hosted server with no setup (keyless free tier)
Where can I get firecrawl-mcp-server?
The source code and installation instructions are on GitHub: https://github.com/firecrawl/firecrawl-mcp-server

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects