Natural-language playlists
Describe a mood, activity, or scene and Kuroma assembles a playlist from your library and beyond.
Product
What Kuroma does for you
Kuroma sits between you and your streaming service, translating messy, human intent into playlists that feel hand-curated but update automatically as your taste and habits evolve.
Adaptive learning
Every skip, replay, and save feeds back into Kuroma’s model, so playlists improve over time.
Cross-session memory
Kuroma remembers what worked last week at 2am vs what you play on commutes, and adapts recommendations accordingly.
Fine-grained controls
Lock in tracks you love, exclude artists or genres, and steer the system without doing manual curation.
Context-aware sequencing
Not just ‘what’ to play, but ‘when’. Energy and tempo are arranged to match the arc of a session.
Platform-friendly design
Built to integrate with existing streaming ecosystems instead of trying to replace them.
How it works
- 1. Intent capture
You describe what you want ("rainy night code session", "train ride at sunrise", "angry cardio").
- 2. Semantic search
Kuroma searches across tracks using embeddings, mood tags, and your past behavior to assemble candidates.
- 3. Sequencing engine
Tracks are ordered to create a progression in energy and color, rather than a random shuffle.
- 4. Live adaptation
As you interact with the playlist, Kuroma adjusts weights in real time and updates future sessions.
Designed for serious listeners
Kuroma is for people who care about sequences, not just songs: DJs, producers, long-form listeners, people who work or study to music.
Instead of fighting the algorithm, you point it where you want to go and let Kuroma handle the curation overhead.