The best music you haven't heard yet: Discovarr helps you find new music that is sonically similar to what you already like.
Pick a track, album, or artist.
Discovarr measures how close songs sound to each other and ranks finds by that distance.
Come back later and pick up from any step.
Pick a seed track and dig a few layers deep, right in your browser. No account and no install.
Discovarr is in alpha on iOS for select music libraries, with more coming soon.
If you self-host a music server, Discovarr bridges the gap between your library and the best music you haven't heard yet. Discovarr knows which tracks are in your library, and which aren't. When you find something you like, add it to your library via Lidarr.
Discovarr is inspired by other great self-hosted music projects, like AudioMuse-AI, Aurral, Explo, and Digarr, and wouldn’t be possible without their foundational work. Discovarr uses the same distilled CLAP model as AudioMuse-AI, so it hears your music the way AudioMuse-AI does.
| Project | Picks music by | Where it stops |
|---|---|---|
| AudioMuse-AI | Picks music by: The sound of your own tracks | Where it stops: Works only with the library you own. |
| Explo | Picks music by: ListenBrainz recommendations from your listening history | Where it stops: Follows listening data, not sound. |
| Digarr | Picks music by: Your listening history or a mood | Where it stops: Suggests artists and albums. Sound matching is optional and experimental. |
| Aurral | Picks music by: Artist recommendations from your listening history | Where it stops: Discovery works at the artist level. |
Bridges your library and music you don't own. Finds can be owned or not owned.
Picks by sonic similarity. It ranks by the audio itself.
Works at any level. Dig from a track, an album, or an artist.
The backend will be free and open source (AGPL) soon.
Sign up with your Apple ID email, and you’ll get a TestFlight invite when we launch.
Sign up for an inviteQuestions? Email hello@discovarr.io