Comment by perrym137

20 hours ago

https://unheard.fm

I was tired of repeat, sponsored, and "safe" music suggestions from Spotify, so I built a discovery engine that puts the control back in the user's hands.

The core idea is simple: You define a "Discovery Model" with explicit constraints (specific genres, release years, track popularity, etc.). The app then uses this blueprint to source tracks.

The results are fresh for two reasons:

- "Known" Track filtering: Excludes all songs saved in your library and recent listening history.

- Active Curation: Uses your custom model, not a vague, opaque algorithm.

It’s built with a local-first mentality and a focus on privacy. No black-box AI "vibe" mixes, just pure, objective discovery based on your criteria.

Hope ya'll find some new gems!

I canceled my Spotify subscription because it would not let me "reset" the algorithm to get fresh suggestions.

A "discovery algorithm" that I used (works great for jazz) consisted on looking up which musicians played on an album that I liked on discogs and searching for more albums from them.

  • Annoying right?

    I am also a huge Discogs fan and unheard.fm actually leverages their APIs to aid with discovery ;)