Firecrawl CLI

Installation & Auth

npm install -g firecrawl-cli
firecrawl login --browser        # Recommended for agents
# or
export FIRECRAWL_API_KEY=fc-YOUR-KEY
firecrawl --status               # Verify: shows credits + concurrency

Commands Summary

| Command | Purpose | |---------|---------| | firecrawl scrape <url> | Scrape single URL | | firecrawl search "<query>" | Web search (+ optional scrape) | | firecrawl map <url> | Discover all URLs on a site | | firecrawl crawl <url> | Crawl entire website (async job) | | firecrawl browser | Cloud browser sandbox automation | | firecrawl agent "<prompt>" | NL-driven web agent queries |

Key Patterns

Scrape (most common):

firecrawl https://example.com --only-main-content          # Clean markdown
firecrawl https://example.com --format markdown,links      # Multiple formats → JSON
firecrawl https://example.com -o output.md                 # Save to file

Crawl a docs site:

firecrawl crawl https://docs.example.com --limit 50 --max-depth 2 --wait --progress -o docs.json

Browser automation (AI agents):

firecrawl browser launch-session
firecrawl browser execute "open https://example.com"
firecrawl browser execute "snapshot"     # Returns @ref IDs
firecrawl browser execute "click @e5"
firecrawl browser execute "scrape"
firecrawl browser close

AI agent query:

firecrawl agent "Find top 5 AI startups and funding" --wait
firecrawl agent "Compare pricing" --urls https://a.com,https://b.com --wait

Full Reference

See references/commands.md for all commands, options, and examples.

Tips

• Use --only-main-content for clean article content (removes nav/footer)
• crawl returns a job ID immediately — use --wait to block or poll with job ID later
• Browser execute default mode is agent-browser (bash) — 40+ commands, best for agents
• spark-1-mini (default) is 60% cheaper than spark-1-pro for agent queries
• Check concurrency limit with --status before parallelizing scrape jobs
• Self-hosted instances skip API key auth automatically when --api-url is set