Feedstock
NPM Package
A web crawler and scraper for TypeScript on Bun and Playwright. It tries a plain HTTP fetch first and only launches a browser when the page needs one.
TypeScript
Bun
Playwright
Web Scraping
CLI
Open Source
Key features
- Fetches over plain HTTP first and moves to a browser only for an empty app shell or a blocked request
- Deep crawling with breadth-first, depth-first, and best-first strategies
- Returns only the formats you ask for, from markdown, HTML, links, images, screenshot, PDF, and structured JSON
- Caches by configuration, with a two-day default and ETag support
- Follows robots.txt, crawl-delay, and Retry-After, with per-domain rate limits
- Runs on Playwright, any CDP browser service, or Lightpanda
- A library and a command-line tool
Feedstock is a web crawler and scraper for TypeScript. It doesn’t start a browser unless it has to. It fetches the page over HTTP, and only if that returns an empty app shell or a block does it open a real browser.
What you get back
You ask for formats and get only those. Markdown comes with real tables, absolute URLs, and citations that aren’t repeated. There is also a compact accessibility snapshot, made for agents that need to refer to elements on the page.
Install
bun add feedstock
bunx playwright install chromiumThe docs are at feedstockai.com and the code is on GitHub.