Momixa uses a variation of the Word2Vec algorithm to learn how songs relate to each other. Thousands of popular Spotify playlists are fed into a neural network that groups songs that frequently appear together. From this neural network, the "coordinates" of each song is extracted.
When you enter two songs, Momixa finds those two songs in its map of the music universe, charts a path between them, and finds songs close that path. This path creates a playlist that transitions between the two songs, sometimes with serendipitous (or wacky) stops along the way.
We are a small team of two developers and a designer with day jobs. We built Momixa because we love listening to and discovering new music!