Hey everyone, I built Feedstock because I kept reaching for crawl4ai for LLM data pipelines but work mostly in TypeScript, and bouncing between languages got old. A few things I focused on: - Speed. Runs on Bun with a fetch-first engine that only spins up a browser when a page actually needs it. - AI-friendly output. Clean markdown, accessibility snapshots that give LLMs a compact view of the page, and change tracking so you only reprocess what's new. - Flexibility. Playwright, CDP, or Lightpanda depending on the tradeoff you want. The usual crawler stuff done well: proxy rotation, rate limiting, robots.txt, anti-bot detection, SQLite caching, and deep crawling with BFS/DFS/BestFirst. It's open source. Would love feedback, especially from folks building data pipelines or agents that need fresh web data.

Feedstock

Crawl, scrape, and extract structured data in TypeScript

Crawl, scrape, and extract structured data in TypeScript