What I like best is how seamlessly Firecrawl fits into my AI workflow. Setup took minutes: I added it as an OAuth connector in Claude, with no API keys to manage, and since then every session can search the web, scrape pages and look up research papers (Integrations, Onboarding). The output is clean, LLM-ready Markdown, and the newer features like structured extraction with a prompt and the Alexandria data providers mean the agent often gets exactly the fields it needs instead of a whole page (AI). It’s fast and reliable, including JavaScript-heavy sites where simpler scrapers fail, and the Usage page showed a 99.6 % success rate across my jobs (Performance). For bigger jobs I use the API directly. Archiving a whole university website cost exactly one credit per page, which made the cost easy to predict. The free tier with 1,000 monthly credits is enough to try it properly before paying (Pricing). The dashboard is clean and the playground lets me test a scrape before automating it (UI). Review collected by and hosted on G2.com.
The dashboard overview isn’t always reliable: the “Credits used – last 7 days” tile showed 0 while the Usage page listed 883 credits used in two days. Credit costs also vary by feature (a search costs 2 credits, a structured extraction 5), so when an AI agent decides on its own what to call, credits can run out faster than expected. A per-source breakdown or a usage alert would help. Review collected by and hosted on G2.com.