▶ Press play. Thirty real minutes the app built. Up fast and stays there. Every control works, so skip ahead with drop › or voice ›.
Pick a mood, set the length, press build. It cuts the music and drops the voices in itself, so what comes out is one continuous mix rather than a playlist you keep skipping through with wet hands.
The voices are the boost. Real people who did the work, landing on the drop, at the point in a set where you were going to stop. You do not have to feel like training before you start.
Six moods, any length from five minutes to two hours, and control over how much talking there is and how far forward the music sits. Say who you want to hear more of, and who you never want to hear again.
That last number counts the music alone, and it is the plain arithmetic of choosing nine tracks out of 102: 102 choose 9 = 2,290,415,157,800. Nine because a thirty minute session runs about nine tracks. It is the conservative count, too, since it ignores the order they play in and it ignores the voices entirely: which of the 594 clips land on top, and where, multiplies it again by a number not worth writing down.
A track plays twelve seconds from one of three places, its calmest stretch, its build or its hardest drop, and a different one each time you press it. A person plays every clip they have, whole, starting with the ones the engine picked for the demos above. Play all shuffles either rack, plays only what your filters left, and comes back to a different opening next visit.