Firecrawl — Claude Code MCP server: what it does, how to install
Firecrawl connects MCP clients to live web search, Markdown scraping, document parsing, and site crawling.
no license updated 2026-10-29
Open on GitHubWhat Firecrawl does
The server converts live web pages into Markdown or schema-validated JSON. It provides firecrawl_scrape for single URLs, firecrawl_search to query web or developer sources, and firecrawl_crawl to traverse multiple pages under a target path.
Agents can navigate dynamic sites using firecrawl_interact to click, submit forms, or handle state. Built-in tools parse PDF files, search research papers on arXiv, and track scheduled content diffs through firecrawl_monitor_*.
The hosted endpoint without an API key offers only three basic tools: search, scrape, and parse. Full crawling, site mapping via firecrawl_map, and higher rate limits require a Firecrawl account and API key.
What it helps with
- extract clean Markdown or JSON from target web pages
- search developer documentation and scientific research papers
- crawl a domain to gather content across multiple paths
- automate browser clicks and form submissions on dynamic pages
Example prompts for the agent
- Search for recent papers on arXiv using
firecrawl_search - Extract documentation pages into Markdown using
firecrawl_scrape - Map all internal URLs for a documentation site via
firecrawl_map
How to install Firecrawl
Install
The MCP registry has no launch command for it: see the repository README
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
How to install Firecrawl in Claude Code?
Run the server locally with env FIRECRAWL_API_KEY=fc-YOUR_API_KEY npx -y firecrawl-mcp. Alternatively, configure your client to connect to the remote endpoint https://mcp.firecrawl.dev/v2/mcp.
What does Firecrawl do?
It gives AI clients tools to search the web, scrape clean Markdown or JSON, parse documents, and crawl whole domains. It also supports browser actions like clicking buttons and filling input fields.
Does Firecrawl require an API key?
The keyless hosted endpoint at https://mcp.firecrawl.dev/v2/mcp provides rate-limited scraping, searching, and parsing without credentials. Advanced operations like firecrawl_crawl and firecrawl_map require a Firecrawl API key.