Recommendarr
LLM-driven movie and TV recommendation web app that uses Sonarr/Radarr libraries and Plex/Jellyfin watch history to generate personalized suggestions.
Recommendarr is a web application that generates personalized movie and TV show recommendations using data from your existing media library and watch history. It integrates with popular media managers and can use cloud or local LLM providers to tailor suggestions to your preferences.
Key Features
- AI-powered recommendations based on Radarr and Sonarr libraries
- Watch history analysis via Plex and Jellyfin, with optional Tautulli and Trakt integration
- Supports multiple AI backends, including OpenAI-compatible APIs and local LLMs
- Web UI with configurable recommendation settings (count and model parameters)
- Light/dark theme support and poster display with fallbacks
- Built-in authentication with optional OAuth login support
Use Cases
- Discover new movies and series that match your existing collection
- Generate recommendations based on what household members actually watch
- Run a local-LLM recommendation workflow for a privacy-focused media setup
Limitations and Considerations
- Recommendation quality depends heavily on the completeness of your library metadata and watch history
- External access should be deployed behind a properly configured reverse proxy and authentication
Recommendarr is a practical companion for Sonarr/Radarr-centric media stacks, combining library context with LLMs to produce tailored suggestions. It fits well in Plex or Jellyfin environments where you want recommendations driven by your own viewing habits.