Sonik: Music Discovery

Most music apps recommend from genre tags, charts, and playlists someone typed by hand. Sonik listens to the actual audio.

Every 30-second preview is run through a neural network (CLAP) that turns raw sound — timbre, rhythm, energy, mood — into a 512-dimensional fingerprint. Songs that genuinely sound alike land close together, so your recommendations come from how music feels, not how it was labeled.

HOW IT WORKS • Swipe right on what you love, left on what you don't. • Sonik builds a personal taste model from your swipes — and trains a neural collaborative-filtering network that learns the patterns behind your choices. • A progress bar shows how far along your personal model is. The more you swipe, the more your feed is driven by your own neural fingerprint instead of generic popularity.

DISCOVER • An endless, personalized deck of new tracks with instant 30-second previews. • Search any genre or artist to steer your feed.

CONNECT • Compare taste with friends and see what they're into. • Share songs directly, and join a community feed to talk music, discover new creators, or produce something yourself.

Sonik is built by a solo developer who thinks discovery should be about the sound itself. No tastemaker gatekeeping, no recycled charts — just your ears and an algorithm that listens.

Free to use. Questions or ideas? soniklegalinquiries@gmail.com