✨ Added
- LLM Provider Integration: Added support for connecting external AI providers such as ChatGPT (OpenAI), Ollama, OpenRouter, LiteLLM or other OpenAI-compatible endpoints.
- AI-Powered Recommendations (Optional): Introduced an advanced recommendation engine that leverages configured LLM providers to generate contextual suggestions based on watch history.
- AI Rationale Display: Each AI-generated recommendation now includes a clear "AI Reasoning" section explaining why the content was suggested.
- Multi-Job Similarity Search: Implemented multi-job searching for content similar to previously watched media or based on selected criteria.
- Sonarr/Radarr Download Profile Sync: Added support for importing and distinguishing download profiles directly from Sonarr and Radarr.
- Enhanced AI Provider Configuration: Expanded AI provider settings with additional input fields, improved visibility, and support for custom base URLs.
- Granular WebUI Settings Management: Settings can now be edited more granularly directly from the WebUI.
- Cron Preset Validation: Added validation for cron presets in AdditionalSettings and SettingsGeneral components.
🛠 Improved
- Recommendation Filtering: Implemented strict duplicate detection to prevent already watched or existing content from being recommended.
- Metadata Handling: Improved TMDb ID matching flow and metadata preservation for Plex and Jellyfin/Jellyseer integrations.
- Configuration Persistence: Enhanced advanced settings persistence and updated configuration parameters for better stability.
- Code Quality: Improved type-hinting and internal structure for better maintainability.
🐛 Fixed
- Malformed Requests Handling: Improved handling of malformed requests to prevent unexpected failures.
- Missing TMDb Metadata: Fixed issues related to incomplete or missing TMDb metadata.
- Advanced Settings Persistence: Resolved issues where beta/advanced settings were not properly saved.
📌 Notes
This release introduces a major functional upgrade with optional LLM-based recommendations and contextual reasoning, while maintaining full backward compatibility.
The focus of this update is:
- Smarter and more transparent recommendations
- Better integration with Sonarr/Radarr and media servers
- Improved configuration flexibility
- Increased stability and reliability
LLM integration is fully optional and requires configuring a compatible provider in Settings.
Feedback and bug reports are always welcome ❤️