Firecrawl Plugin for Claude Code

Turn any website into clean, LLM-ready markdown or structured data with Firecrawl.

Features

• Scrape: Extract content from any webpage as clean markdown
• Crawl: Automatically discover and extract content from entire websites
• Search: Search the web and get scraped results
• Map: Discover all URLs on a website

All features include automatic JavaScript rendering, anti-bot handling, and proxy rotation.

Installation

1. Install the Plugin

/plugin install firecrawl@claude-plugins-official

Important: After installing, restart Claude Code to load the Firecrawl MCP server.

2. Set Up Your API Key

Run /firecrawl:setup to configure your API key. It will detect where the plugin is installed and guide you.

Or manually add your API key using one of these methods:

Option A: Claude Code Settings (Recommended)

Add to the same settings file where the plugin is enabled:

{
  "env": {
    "FIRECRAWL_API_KEY": "fc-YOUR-API-KEY"
  }
}

File locations by scope:

• ~/.claude/settings.json - Global (all projects)
• .claude/settings.json - Project (shared with team)
• .claude/settings.local.json - Local (gitignored, personal)

Option B: Shell Profile

# Add to ~/.zshrc or ~/.bashrc
export FIRECRAWL_API_KEY=fc-YOUR-API-KEY

Note: Running export directly in terminal only lasts for that session. Add it to your profile file for persistence.

Get your free API key at: https://firecrawl.dev/app/api-keys

3. Verify Setup

Restart Claude Code, then run /firecrawl:setup to confirm everything is working.

Usage

Slash Commands

Command	Description
/firecrawl:setup	Configure API key and verify setup
/firecrawl:scrape	Scrape a single webpage to markdown
/firecrawl:crawl	Crawl an entire website
/firecrawl:search	Search the web and scrape results
/firecrawl:map	Discover all URLs on a website

Available Formats

When scraping, you can request different output formats:

• markdown - Clean markdown content (default)
• html - Raw HTML content
• screenshot - Screenshot of the page
• links - Extract all links
• summary - AI-generated summary

Direct Tool Usage

The plugin exposes Firecrawl's MCP tools directly:

• firecrawl_scrape - Single page scraping
• firecrawl_batch_scrape - Multiple URL scraping
• firecrawl_crawl - Website crawling
• firecrawl_check_crawl_status - Check crawl progress
• firecrawl_map - URL discovery
• firecrawl_search - Web search with scraping

Examples

Scrape a documentation page:

Scrape https://docs.firecrawl.dev/introduction and summarize the key points

Get a summary of a page:

Scrape https://firecrawl.dev with summary format

Research a topic:

Search for "best practices for React testing" and compile the key recommendations

Crawl a documentation site:

Crawl the entire documentation at https://docs.firecrawl.dev and create a summary

Discover site structure:

Map all URLs on https://firecrawl.dev

Configuration

Environment Variables

Variable	Required	Description
FIRECRAWL_API_KEY	Yes	Your Firecrawl API key
FIRECRAWL_API_URL	No	Custom API endpoint (for self-hosted)
FIRECRAWL_RETRY_MAX_ATTEMPTS	No	Max retry attempts (default: 3)
FIRECRAWL_RETRY_INITIAL_DELAY	No	Initial retry delay in ms (default: 1000)
FIRECRAWL_CREDIT_WARNING_THRESHOLD	No	Credit warning threshold (default: 1000)

Self-Hosted Deployment

Firecrawl can be self-hosted. Set FIRECRAWL_API_URL to point to your instance:

export FIRECRAWL_API_URL=https://your-firecrawl-instance.com

See Firecrawl documentation for self-hosting instructions.

Resources

• Firecrawl Documentation
• API Reference
• GitHub Repository
• Get API Key

License

This plugin is licensed under AGPL-3.0, consistent with Firecrawl's open-source license.

Support

• Firecrawl Discord
• GitHub Issues
• Email: [email protected]