Skip to main content
MCP server CX Codex

Firecrawl MCP Server — Claude Code MCP server: setup & tools

Extracts clean Markdown from web pages, crawls whole domains, and searches technical or academic papers. Built for AI agents needing live web context.

Install into Claude Code · Platform CLI: remote server All methods and platforms ↓
claude mcp add --transport http firecrawl-mcp-server https://mcp.firecrawl.dev/v2/mcp
Open on GitHub

What Firecrawl MCP Server does

The server bridges MCP-compatible clients with the Firecrawl API. It delivers tools to scrape individual URLs via firecrawl_scrape, discover links using firecrawl_map, crawl subpages with firecrawl_crawl, and parse PDF documents. It also queries dedicated developer indexes such as GitHub issues, government databases, and scientific repositories including arXiv and PubMed.

For dynamic interfaces, firecrawl_interact executes browser actions like clicking, typing, and form submissions. Each call completes only a single turn before yielding control. Step-by-step interactive sessions within a single prompt are unsupported, though state can persist across calls via scrapeId.

The hosted keyless endpoint exposes only three capabilities: search, scrape, and parse. Accessing deep crawling, URL mapping, and agent research jobs requires supplying a valid FIRECRAWL_API_KEY through environment variables or request headers.

What it helps with

  • scrape web pages and PDF files into clean Markdown or JSON
  • map site structures and extract content across multiple pages
  • search research papers on arXiv, PubMed, and bioRxiv
  • fill forms and click elements on dynamic JavaScript pages

Example prompts for the agent

  • Scrape the API documentation at this URL and format it as Markdown
  • Search arXiv for recent papers on LLM agents using Firecrawl
  • Map all subpages under example.com without fetching their content

How to install Firecrawl MCP Server

1Platform CLI: remote server

claude mcp add --transport http firecrawl-mcp-server https://mcp.firecrawl.dev/v2/mcp

2Platform CLI

claude mcp add --env FIRECRAWL_API_KEY=<FIRECRAWL_API_KEY> --transport stdio firecrawl-mcp-server -- npx -y firecrawl-mcp

Replace <…> with your own values

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Firecrawl MCP Server

How to install Firecrawl MCP Server in Claude Code?

For hosted access, run claude mcp add --transport http firecrawl-mcp-server https://mcp.firecrawl.dev/v2/mcp. To run locally via stdio with full tooling, run claude mcp add --env FIRECRAWL_API_KEY=<FIRECRAWL_API_KEY> --transport stdio firecrawl-mcp-server -- npx -y firecrawl-mcp.

What does Firecrawl MCP Server do?

It equips assistants with tools to search the live web, crawl multi-page sites, extract clean Markdown from web pages, and interact with browser elements.

Does Firecrawl MCP Server require an API key?

The hosted keyless endpoint functions without an API key but limits tools to search, scrape, and parse. Advanced tools like crawl, map, and autonomous research jobs strictly require FIRECRAWL_API_KEY.

Similar MCP servers

All MCP servers for Browser and scraping →

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor