Scraping Kayak for live flight prices needs --wait-for 15000
firecrawl flight-scraping
Kayak/Google Flights load results client-side, so a plain firecrawl scrape returns only the page chrome (search box + footer, no flight list). Adding --wait-for 15000 gets the rendered results. Google Flights URLs without proper tfs= base64 tokens silently render the homepage with unrelated cheap-flight teasers. Kayak URLs like /flights/DEL-SFO,SJC,OAK/2026-06-15?sort=price_a&fs=stops=~1 work directly and support multi-airport destinations comma-separated. Parse the markdown by splitting on the airline-logo image regex ( — each split block contains exactly one flight card with airline + stop + layover + price.
For aggregator scrapes: --wait-for 15000 always; prefer Kayak URL format over Google Flights; split-on-airline-logo to parse cards. Skip firecrawl agent for tabular scrapes — it took 6min vs 15s for direct scrape and wrote no useful output.