Firecrawl
Open-source URL → LLM-ready markdown converter. 97,000 GitHub stars, Y Combinator-backed. Per the source: passed Scrapy’s star count, the long-time default Python web scraper. “If your agent needs to read the internet, this is the standard.”
What it does
- Point at any URL → get clean markdown or structured data
- No HTML parsing or post-processing required
- Has its own API server for direct LLM/agent integration
- Plugs directly into Claude Code
Why it matters
Scrapy passed-by-star-count is a meaningful inflection point — it suggests agent-native scraping has overtaken the legacy Python scraping ecosystem in active interest. Firecrawl is to LLM-era data extraction what Scrapy was to the structured-data era.
The wiki has tracked sister tools — crawl4ai, stagehand — but Firecrawl has the largest mindshare of any of them per the source.
Type and platform
- Type: CLI + API server
- Platforms: model-agnostic (any agent or LLM)
- Status: tracked
Hands-on notes
(empty)
Sources
- The Only Claude Skills You Need in 2026 (2026-04-08)