Hey Product Hunt I built MatchBeats because I kept getting frustrated with music discovery being purely algorithm-driven. Apple Music and Spotify can tell you “here’s a song similar to what you listen to,” but they can’t tell you “here’s a person whose taste is basically identical to yours.” That felt like a missing piece, music discovery has always worked best socially, through friends with great taste, not just a recommendation engine guessing at genre tags. The problem I was actually trying to solve: how do you quantify “similar taste” in a way that’s more meaningful than shared genres? My approach evolved a lot here, I started with a simple genre-overlap score, but that felt too blunt (two people can both like “rock” and have nothing in common). I ended up building a weighted match algorithm that looks at shared artists and tracks (position-weighted, so your #1 favorite artist matters more than your #50), genre distribution, era/decade overlap, and even how mainstream vs. niche someone’s listening habits are because a mainstream listener and a deep-cuts listener rarely vibe the same way even with genre overlap. The other thing that changed along the way: I originally just wanted to show a match percentage, but early testers kept asking “why did we match?” so I added a full breakdown showing exactly which artists, tracks, and eras drove the score. That transparency turned out to matter more than the score itself. Right now it’s Apple Music only (Spotify’s developer access requirements are a real wall for a small project, happy to talk about that in the comments if anyone’s curious), and still early, so go easy on sparse matches. Would love feedback, especially from anyone who’s tried other “music social” apps and can tell me what’s missing.

MatchBeats

Music discovery powered by people, not algorithms

Music discovery powered by people, not algorithms