I kept rebuilding web scraping infrastructure for AI projects, so I open-sourced it.

Two functions:

- `scrape()` – any URL to markdown

- `crawl()` – entire sites with depth/page limits

Self-hosting:

- Docker image available

- Manages its own browser pool

- Proxy support built-in

- No external dependencies

Deployment guide: https://docs.reader.dev/documentation/guides/deployment

GitHub: https://github.com/vakra-dev/reader

Built with TypeScript, runs anywhere you can run Docker. Happy to answer questions about the architecture or deployment.

https://blog.cloudflare.com/markdown-for-agents/ https://news.ycombinator.com/item?id=46997526