Sagarr is an AI-powered media recommendation engine designed to integrate seamlessly with the "Arr" stack (Sonarr, Radarr, Overseerr) and Plex/Tautulli. It acts as a personalized discovery layer, using your watch history to generate dynamic, creative recommendations.
⚠️ Status: Early Beta / Not Production-Hardened
Sagarr is experimental software. It has not undergone a formal security review and is not intended for multi-tenant or high-risk environments.
Use at your own risk, and only expose it to the internet behind a trusted reverse proxy, with strong secrets and access limited to users you trust.
- AI-Powered Recommendations: Uses LLMs (OpenAI, Ollama, etc.) to invent creative categories (e.g., "Gritty 80s Cyberpunk") based on your taste.
- Plex Integration: Sign in with your Plex account.
- Tautulli Sync: Automatically ingests your watch history to understand your preferences.
- Overseerr Integration: Checks availability of recommendations and allows one-click requests.
- "Seen It" Tracking: Mark items as seen and rate them to refine future suggestions.
- Modern UI: A responsive, dark-mode interface built with React and Vite.
See docs/setup.md for detailed installation instructions using Docker.
- Backend: Python (FastAPI), SQLAlchemy, SQLite
- Frontend: React, Vite
- AI: OpenAI API / Generic OpenAI-compatible endpoints
- Containerization: Docker, Docker Compose
MIT