Skip to content

v0.1.0 - Initial Beta Release

Choose a tag to compare

View all tags
@rdpharr rdpharr released this 26 Dec 21:36
· 43 commits to main since this release
v0.1.0
926c262

Privacy-First Personalized Recommendations for Jellyfin

This is the initial beta release of Local Recommendations, a Jellyfin plugin that generates personalized movie and TV show recommendations based entirely on your local watch history.

Key Features

  • Per-User Personalization: Each user receives recommendations tailored to their viewing history
  • Content-Based Filtering: Uses TF-IDF embeddings and cosine similarity to find similar content
  • Virtual Library Integration: Recommendations appear as dedicated libraries accessible from all Jellyfin clients (web, mobile, Roku, etc.)
  • Privacy-First Design: All processing happens locally on your server with zero external dependencies or tracking
  • Configurable Weighting: Favorite boost (2.0x), rewatch boost (1.5x), recency decay (365-day half-life)
  • Smart Filtering: Abandoned series exclusion, minimum watch history requirement
  • Flexible Updates: Daily scheduled task with manual refresh option
  • Performance Optimized: Handles libraries of 2,000+ items efficiently

Requirements

  • Jellyfin Server: 10.11.5 or later
  • .NET Runtime: 9.0
  • Target ABI: 10.11.0.0

Installation

See the README for detailed installation and setup instructions.

Documentation

Known Limitations

  • Requires manual one-time library setup per user (5-10 minutes)
  • Cold start: Users with fewer than 3 watched items receive popular content recommendations
  • No collaborative filtering (recommendations based solely on individual user's history)

Support

Changelog

See CHANGELOG.md for detailed release notes.

Assets 5
Loading