Comment by shadeslayer

It currently seems to use Spotify's audio analysis API [1], which might be hard to replace.

[1] https://developer.spotify.com/documentation/web-api/referenc...