Privacy
Learnarr runs on your server and keeps everything in its /config folder. There's no account with us, no analytics and no telemetry. The project never sees your data.
What leaves your server, and when:
| Goes to | What | When |
|---|---|---|
| Your AI provider | An episode's subtitle text, or one line | When you ask for suggestions, or make a card from a tapped word |
| OpenSubtitles.com | The film or episode's title and IDs, your OpenSubtitles login | When the media server has no subtitle and you ask Learnarr to download one |
| Wiktionary | The word you tapped | When you tap a word and open its definition |
| lindat.mff.cuni.cz | Nothing about you: a model file is downloaded | Once per language |
| Your push service or ntfy server | The notification text | When you've set up notifications |
With a local AI model (Ollama, LM Studio), subtitle text never leaves your network.
Passwords are hashed. The media server token, AI key and OpenSubtitles passwords are encrypted at rest with a key generated in /config/encryption.key. Keep that file with your backups: without it, those secrets can't be read and have to be entered again.
Subtitles are stored per person and never shared, even inside the household, because a file downloaded with one person's OpenSubtitles.com account is licensed to them. Learnarr never serves subtitle text to anyone but the person who fetched it.