AI workflow
RSS pipeline
Some of the content I want to follow doesn’t publish a feed anymore. RSS Service fixes that. Give it a URL, and it builds the feed for you, whether that’s a plain website, a YouTube channel or playlist, or a Reddit subreddit or user page. It figures out which kind of source you pointed it at and applies the right extraction logic automatically, so you’re not picking a “mode” by hand.
Under the hood, it renders JavaScript-heavy pages with a headless browser, so single-page sites that normally defeat scrapers still produce usable items. Every item gets a SHA-256 GUID for deduplication, which keeps the same post from showing up twice when a page reshuffles. Feeds come out as standard RSS 2.0 or Atom, so anything from a phone app to a homelab reader can consume them.
The dashboard is where you manage everything. Each feed shows its platform, item count, and refresh status. You can preview what a feed will contain before you save it, edit a feed’s config with automatic re-detection, and trigger a manual refresh whenever you want. Auto-refresh is configurable from every 15 minutes up to once a day. Feed configs export to JSON for backup, individual feeds download as static XML, and the whole UI works on mobile with a collapsible sidebar plus light and dark themes.
It’s built to self-host. The stack is Node.js and Express with TypeScript, PostgreSQL for storage, and Tailwind CSS v4 with daisyUI on the front end. It ships as a Docker Compose setup with PostgreSQL and Chromium bundled in, designed to sit behind a reverse proxy like Caddy or nginx. Database migrations run automatically on startup. Reddit feeds need no authentication at all. YouTube feeds require a free API key from the Google Cloud Console. MIT licensed.
