fit filter — Crawl4AI sám ořeže menu/footer/reklamy.
Pro batch více URL použij API POST /scrape/crawl.
Stejný backend bez UI — kdykoli z agenta / skriptu:
curl -X POST https://router.czechai.io/scrape/md \
-H 'Content-Type: application/json' \
-d '{"url":"https://sreality.cz/...","filter":"fit"}'
# Crawl celý web (max 100 stránek)
curl -X POST https://router.czechai.io/scrape/crawl \
-H 'Content-Type: application/json' \
-d '{"url":"https://example.com","max_depth":2,"max_pages":50}'
# Schema extrakce přes naše LLM (GPU 170)
curl -X POST https://router.czechai.io/scrape/extract \
-H 'Content-Type: application/json' \
-d '{"url":"https://...","schema":{"price":"number","title":"string"}}'
unclecode/crawl4ai:latest @ server 117:11235 ·
Reverse proxy: 179 · LLM: Qwen 3.6 35B Claude-4.7 distilled (GPU 170)