OfficialFeaturedMCP server · stdioRequires credentials

🔥Firecrawl

Give Claude web scraping, crawling, search, and structured extraction via the Firecrawl API.

What it does

Firecrawl's official MCP server gives Claude reliable web scraping and crawling that returns clean, LLM-ready content.

What it gives Claude:

Scrape any page to markdown or structured JSON, with batch scraping for many URLs
Crawl whole sites asynchronously and map all indexed URLs
Web search with optional inline content extraction
LLM-powered structured data extraction from pages
Autonomous research agent for multi-step web queries
Works against Firecrawl Cloud or a self-hosted instance

Needs a Firecrawl API key (free tier available) set as FIRECRAWL_API_KEY. Maintained by Firecrawl (Mendable).

Before you start

Get a Firecrawl API key (free tier available) at firecrawl.dev/app/api-keys. Swap YOUR_API_KEY in the command for it.

1
Make sure Claude is on your device and in your terminal.
MCP servers register with Claude Code via claude mcp add. You need Claude Code installed first:
```
npm i -g @anthropic-ai/claude-code
```
2
Set up the server on your machine.
Setup
Sign up to copy
```
claude mcp a•• ••••••••• •• •••• ••••• •••••••••••••••••••••••••••••• •• ••• •• •••••••••••••
```
Sign up free to reveal
Runs the project's installer (clone + dependencies + config). Follow any prompts. After this finishes, the server is ready to register with Claude Code in the next step.
Configure credentials.
Requires a Firecrawl API key (FIRECRAWL_API_KEY), created free at firecrawl.dev/app/api-keys.
Restart Claude Code.
Quit and reopen Claude Code so it picks up the new MCP registration. Run /mcp inside Claude Code to confirm the server is connected.
Use it.
Ask Claude naturally. It will call the server's tools when your request matches one. Tool calls show up inline so you can see exactly what Claude is doing.

Prefer to read the source first? Open on GitHub.