Open tests of web scraping and crawling APIs. Same inputs for every tool, scoring you can check, and our own numbers included, wins and losses.
7 APIs scraped the same 1,000 URLs from Firecrawl's public dataset. We scored whether each returned the main content, how clean the markdown was, and how long it took.
Every tool gets the same URLs in the same order with default settings and caching off. No per-tool tuning and no dropping the pages a tool struggles with.
We use public datasets and published scoring rules, not a model judge. Every metric is defined in full, with the scoring code, on its page.
We build WebCrawlerAPI and it appears in these tests. Results are published as measured and dated, including the metrics where we lose.
A benchmark is a starting point. Scrape your own URLs and compare the markdown yourself. No credit card needed.