The data sources that power your recommendations
Watch Tonight works by combining three layers of data about your film taste — and the more layers you provide, the better the results.
### 1. Your media library
If you connect Plex, Jellyfin, or Emby, KinoChain indexes your library and knows exactly what films you own. This is the pool from which recommendations are drawn — films you can watch right now, without renting or subscribing to anything new.
### 2. Your watchlist
Your Plex watchlist, Letterboxd watchlist, and KinoChain watchlist all feed into the same pipeline. Films you\'ve flagged as "want to watch" are prioritised when they match your taste profile.
### 3. Your watch history
This is the most important layer. Every film you rate, review, and log teaches the system about your taste. The algorithm looks at directors, genres, themes, and patterns in what you've loved — then finds matches in what you own.
### 1. Your media library
If you connect Plex, Jellyfin, or Emby, KinoChain indexes your library and knows exactly what films you own. This is the pool from which recommendations are drawn — films you can watch right now, without renting or subscribing to anything new.
### 2. Your watchlist
Your Plex watchlist, Letterboxd watchlist, and KinoChain watchlist all feed into the same pipeline. Films you\'ve flagged as "want to watch" are prioritised when they match your taste profile.
### 3. Your watch history
This is the most important layer. Every film you rate, review, and log teaches the system about your taste. The algorithm looks at directors, genres, themes, and patterns in what you've loved — then finds matches in what you own.
How logging improves your results
The recommendation quality curve is steepest in the first 50–100 logs.
1
0–10 films logged
Basic genre-based recommendations. The system knows your general preferences but not the nuances.
2
10–50 films logged
Director affinity emerges. The system notices you favour certain directors and surfaces their other films in your library.
3
50–100 films logged
Deep pattern matching. Themes, eras, cinematographers, and "if you liked X, you'll appreciate Y" chains start surfacing.
4
100+ films logged
The system understands your taste well enough to suggest films outside your usual patterns — the ones that surprise you but don't miss.
What we store and what we don't
Privacy matters, especially when connecting your personal media server.
**What KinoChain stores:**
- Library titles (film name, year, TMDB ID) that you choose to sync
- Watch status (watched/unwatched) if you enable diary sync
- Your Plex/Jellyfin/Emby connection metadata (server URL, token — encrypted at rest)
**What KinoChain never stores:**
- Your media files
- Your server login credentials
- Library contents beyond title metadata
- Any user data from your server beyond what's needed for title matching
You can disconnect any service at any time from Settings → Apps. All synced data is permanently deleted.
**What KinoChain stores:**
- Library titles (film name, year, TMDB ID) that you choose to sync
- Watch status (watched/unwatched) if you enable diary sync
- Your Plex/Jellyfin/Emby connection metadata (server URL, token — encrypted at rest)
**What KinoChain never stores:**
- Your media files
- Your server login credentials
- Library contents beyond title metadata
- Any user data from your server beyond what's needed for title matching
You can disconnect any service at any time from Settings → Apps. All synced data is permanently deleted.
Ready to log more?
Chains are the fastest way to discover and log films connected by directors, actors, or themes. Each chain is a curated path through cinema.