Tend
Library
Scan, read, and organize every track — across FLAC, MP3, AIFF, WAV, AAC, OGG, OPUS. Tags get scored. Gaps get surfaced. Intent gets honored. dhun never modifies your source files unless you say so.
Local-first music platform
melody. lagan.
A library that listens back.
A local-first music platform for serious listeners. Tend your library, and what was hiding in it begins to surface — in playlists you build, in patterns you start to see, in the tracks dhun begins to suggest because it has finally learned what your music means to you.
macOS now · Windows + Linux soon
Capabilities
Five surfaces under one platform. Each does its work; together they make your library legible — to itself, and to you.
Scan, read, and organize every track — across FLAC, MP3, AIFF, WAV, AAC, OGG, OPUS. Tags get scored. Gaps get surfaced. Intent gets honored. dhun never modifies your source files unless you say so.
Manual playlists, smart playlists, eclectic-mode sets that adapt as you build. Export to Rekordbox, Serato, Traktor, M3U. dhun listens to every reorder, keep, and skip — playlists get sharper the longer you use it.
Tracks you've forgotten. Tracks adjacent to what's playing. Tracks that fit a moment you didn't realize you were in. Dig surfaces what's already yours, locally — your library's own intelligence, not someone else's algorithm.
Cards, charts, and contextual insights — your library, mirrored back. The tracks you've played most. The tracks you haven't heard in months. The eras getting heavier in your listening this season. The patterns you didn't realize you had.
Generative art derived from your library's signature. Every playlist, crate, and profile carries a Rangoli mark unique to what it represents — a face for what you've built. A design system in its own right, with room for outside artists to add their own visual languages.
100% local · never uploaded · never analyzed online · never shared
Underneath every other surface, RaagGraph is learning. The relationships between every track in your library — across rhythm, harmony, energy, texture, era, context, and more — compound on your machine, every time you play, move, or tag a track. The longer you use dhun, the more your library knows itself, and the more it stays yours.
What you'll build
Playlists in dhun aren't just lists of files. They're shapes of listening — sets that hold an energy arc, an era, a thread you can follow for weeks. A few of the shapes that show up:
Twenty-five years of chasing sound. Mixtapes, a thousand CDs, the iTunes era, iPods, lost hard drives I still grieve. Streaming and unstreaming. Discovering Asian Underground and falling deeper into electronic music than I knew was possible.
I've listened to mystics share sacred sound in Morocco and heard dub playing outside Bob Marley's home. I ran 30 Days of Sound — a new way of listening every day in my own city. I've built playlists that capture what friendship sounds like, what home sounds like in countries I've travelled through. Cheb i Sabbah taught me how to listen to the world. I carry that forward.
Then one day I looked at my own library — decades of music across drives and formats and services — and knew it deserved better. No algorithm built for someone else's taste was going to help me find my own story inside it.
So I'm building the platform I always needed. One that plays along while I figure out the soundtrack to what's unfolding. One that listens back. That's dhun.
— Ajesh