Feeds.fun is a self-hostable news reader designed for users overwhelmed by information overload. It automatically tags news articles using AI and allows users to define scoring rules based on these tags, enabling personalized filtering and sorting to focus on essential content.

How It Works

The system employs a multi-worker architecture. A "loader" worker fetches and parses RSS feeds, while a "librarian" worker analyzes entries. The librarian utilizes configurable "tag processors," including simple domain extraction, native feed tags, and advanced LLM-based (OpenAI/Gemini) tag generation. Users can define scoring rules based on these tags to prioritize or filter content.

Quick Start & Requirements

• Backend: pip install ffun, ffun migrate, uvicorn ffun.application.application:app --host 0.0.0.0 --port 8000 --workers 1, ffun workers --librarian --loader
• Frontend: npm install feeds-fun, npm run build-only --prefix ./node_modules/feeds-fun, copy dist directory.
• Prerequisites: Python, PostgreSQL, OpenAI or Gemini API key (for full AI features), Node.js/npm for frontend build.
• Docker: ./bin/build-local-containers.sh and docker compose up -d for local development.
• Docs: feeds.fun, blog.feeds.fun

Highlighted Details

• AI-powered automatic tagging via OpenAI/Gemini.
• Configurable tag processors and LLM model settings.
• Multi-user support with SuperTokens authentication.
• Customizable scoring and filtering rules.

Maintenance & Community

The project is primarily maintained by a single developer ("Tiendil"). A roadmap is available.

Licensing & Compatibility

The repository does not explicitly state a license in the README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project is in active development with many features planned. Configuration for LLM processors is complex and requires careful setup of API keys and potentially custom API endpoints. The README notes that backend configuration output is not yet user-friendly.