germondai/trawl ? reverse-engineered prompt

Build me a self hosted scraping service that I can run in Docker and connect to my Prowlarr or Jackett setup as a FlareSolverr replacement.

It should be able to fetch normal pages quickly, and when a site throws a Cloudflare or other anti bot challenge, it should keep trying with a browser based solve flow and captcha handling so the request still succeeds. I want both a simple API endpoint for getting a page and a proxy mode for sites that need the full browser fingerprint to work. Please make it easy to start with docker compose, include a health check, and store session data so repeat requests can reuse solved cookies when possible.

If it helps, look up current docs online for the browser automation and captcha solving pieces. I’d also like clear setup notes for connecting it to the rest of my arr tools and for trusting the local CA if the proxy mode is used.