How it works
From raw archive to prepared library in three steps
Point Metamix at your library
Add your archive folder (or Beets database) and your Rekordbox XML export. Metamix mirrors both into its local database and reads the Rekordbox beatgrid for every track — the timing authority nothing in Metamix overrides. Browse and search work immediately.
Analyze and tag locally
Queue the full library and let it run. Analysis runs on the Neural Engine, tagging on local models — a 20k-track library processes overnight in a resumable queue. Interrupted jobs pick up where they stopped. Mixed In Key is optional; energy and groove are computed without it.
Review, then write back safely
Review proposed cues and tags, adjust anything that isn't right, then approve. Metamix snapshots Rekordbox, shows you the exact diff, applies it, and audits every change. Don't like the result? One click undoes the whole batch.
The safety net
Will it overwrite my Rekordbox edits? No.
Your manual edits are sacred
Tracks where you've made manual Rekordbox edits are detected and blocked from automated writes.
Every write is a diff first
Before anything changes, you see precisely what will be written — every cue, every tag, every field, before and after — and approve it.
Everything is undoable
Snapshots before every write, an append-only audit log after. One click rolls an entire batch back.
Crash-safe by construction
Jobs are persisted before they run and recover cleanly after power loss, crashes, or forced quits.
Bring the workflow to your library
Join the waitlist for early access and the locked-in launch price.
Join the waitlist