Firecrawl — Structured Web Scraping / Crawl / Search
Status: ✅ Live-tested 2026-07-15 (scrape example.com → success, statusCode 200).
Key: FIRECRAWL_API_KEY (env.sh). API: https://api.firecrawl.dev/v1
When to use
- Clean Markdown extraction from any page (better structure than raw web_fetch for complex/JS pages).
- Site crawl (multi-page → structured docs) for research/tenders/competitor scans.
- Web search with structured results as an alternative/complement to web_search + Perplexity.
- Complements our
stealth-browser(anti-bot) andbrowser(full automation): Firecrawl = fast structured extraction; stealth-browser = anti-bot heavy sites; browser = interactive flows.
Usage
. /data/.openclaw/secrets/env.sh
python3 skills/firecrawl/firecrawl.py scrape "<url>" [--format markdown|html|links]
python3 skills/firecrawl/firecrawl.py search "<query>" [--limit 5]
python3 skills/firecrawl/firecrawl.py crawl "<url>" [--limit 10]
Routing (DOMAIN_MODEL_RANKINGS integration)
- Single clean page → Firecrawl scrape (fast, structured).
- Anti-bot / login-walled → stealth-browser.
- Interactive multi-step → browser tool.
- Deep multi-source research → Deep Research skill + Perplexity + Firecrawl search combined.