BlackTape uses MusicBrainz and Discogs — open, community-maintained databases — to index artists and rank them by how unique they are within their genre. The more niche the artist, the higher they surface. It's the   
  inverse of how Spotify's algorithm works.

  I got frustrated watching recommendation algorithms flatten discovery. The same artists keep surfacing out of 10+ million indexed. MusicBrainz has 2.6 million artists catalogued with rich genre tags, scene data, and  
  regional metadata. Discogs has release metadata going back 80 years. Combine those two databases and score artists by distinctiveness rather than popularity, and the discovery space opens up completely.

  What it does:

  - Search by genre/scene with atomic tag combinations

  - Discover feed ranked by uniqueness score (rare = surfaceable)

  - Full artist pages: discography, tags, related artists, scene data

  - Spotify playback integration (optional)

  - Time Machine: browse artists by decade

  - Style Map: visual genre/scene navigation

  - Knowledge Base: genre relationship graph

  No tracking, no platform API dependency for the core discovery data. Desktop app built with Tauri + SvelteKit.

  GitHub: https://github.com/AllTheMachines/BlackTape

  Site: https://blacktape.org

  Happy to talk about the MusicBrainz pipeline, the uniqueness scoring, or the open-data approach to discovery.