Feeds Fun

Feeds Fun is an RSS reader for people who want to prioritize articles across their subscriptions. It assigns tags automatically and uses your scoring rules to decide which entries appear first.

• Tag-based ranking: Create rules that score articles by their tags, then sort your reading list by score.
• Filters and reading history: Include or exclude articles by tag, sort by date, and track what you've already read.
• Automatic tagging: Use LLM analysis and simpler tags such as a site's domain; tag normalization reduces inconsistent labels so you can reuse them in rules and filters.
• Feed sources: Read subscriptions from Reddit, GitHub, YouTube, ArXiv and Hacker News, with source-specific support for discovering feeds and processing article formatting.

You can self-host the reader with Docker for a single person or multiple users. A free hosted service at feeds.fun also provides curated feed collections.

AI tagging supports OpenAI and Google Gemini, whose cloud services require an API key. Other models can connect through OpenAI-compatible or Gemini-compatible APIs, including custom endpoints. The reader runs on your server when self-hosted, while the endpoint you choose determines where article text goes for AI analysis. Fetching feeds requires network access.