| name | firecrawl-cli |
| description | Lightweight Firecrawl web scrape/search/map via pi-firecrawl CLI. Use for web scraping, Firecrawl search, page extraction, sitemap URL discovery, or structured page data when MCP tool schemas are too token-heavy. |
Firecrawl CLI
Use bash CLI. Do not call Firecrawl MCP tools directly unless CLI fails.
Commands
pi-firecrawl scrape <url> --format markdown
pi-firecrawl scrape <url> --format json --prompt 'extract prices' --schema '{"type":"object"}'
pi-firecrawl search '<query>' --limit 5
pi-firecrawl search '<query>' --limit 5 --scrape
pi-firecrawl map <url> --search 'webhook docs' --limit 20
pi-firecrawl extract <url> --prompt 'extract product details' --schema '{"type":"object"}'
Rules
- Specific fields/data →
--format json with schema.
- Full page/article →
--format markdown.
- Unknown URL →
search first, then scrape selected URL.
- SPA empty result → retry
scrape --wait 5000, then map --search, then search.
- Output can be large. Use low
--limit and focused prompts.