FAQ
Frequently asked questions
A Mac with Apple Silicon running macOS 27 or later, and Rekordbox 7. Metamix is a native macOS app — there is no Windows or Linux version and no pre-macOS 27 support. The analysis path uses Apple's MusicUnderstanding frameworks on the Neural Engine.
No. Analysis, tagging, and cue generation all run locally on your Mac. There is no cloud upload, no account, and no listening service. The only outbound connections are update checks and model downloads.
Never. Tracks where you have made manual Rekordbox edits are marked as yours and automated writes to them are blocked. Every automated write is planned as a diff, audited, and reversible before it touches Rekordbox.
Metamix treats Rekordbox's beatgrid as the timing authority. Cues are snapped to the exact grid Rekordbox plays with, so cue points land where the music actually is — not where another analyzer guessed. If a track has no valid grid, Metamix refuses to place final cues rather than write something wrong.
$20 one-time. That includes the full cue and tagging automations for lifetime use and 1 year of updates. No subscription. The library browser and discovery features are free.
Yes. You can scan your library, browse, search, and build discovery crates for free. The paid license unlocks the automated cue engine and tagging.
Metamix is in active development. Join the waitlist for the first look, early access, and the locked-in launch price.
Metamix can read Mixed In Key exports and normalize its cue landmarks onto your Rekordbox grid, so existing MIK work feeds the pipeline instead of fighting it. Mixed In Key is optional — energy, groove, and cue roles are computed locally without requiring a MIK license. Mixed in Key's key detection is not replaced by Metamix; it is used as-is when present.
Yes. Every write is part of an audited batch with a one-click undo, and Rekordbox snapshots are taken before anything changes.
Point Metamix at a Beets library or a plain filesystem archive folder, plus your Rekordbox 7 collection (XML export and master.db). Descriptive metadata comes from the source; performance state stays with Rekordbox. One active source mode per profile.
Rekordbox and CDJs expose four top-level MyTag categories for hardware filtering. Metamix maps them to acoustically determinable mix dimensions — Role, Components, Groove, and Atmosphere. Genre and other tags that don't need hardware buttons live in comments and the Metamix database instead.
No. Core browse, search, analysis, tagging, and cues work offline without third-party accounts. Optional providers (for extra metadata or similarity) are bring-your-own-key and degrade gracefully when missing — Metamix never fabricates a value it couldn't fetch.
No. It builds on Rekordbox. Metamix consumes the beatgrid Rekordbox already trusts for timing, and writes cues and tags back into your performance library under a controlled, audited path — not a parallel re-analysis of the booth stack.
Yes. Metamix is designed to support 5,000–20,000 tracks and aims to feel useful well below that ceiling. Smaller libraries get the same local analysis, browse, and safety model without needing a cloud tier.
Database writes to Rekordbox are blocked and queued while Rekordbox is running. File-tag and source writes can continue. Metamix never sneaks a master.db change under a running app — you gate every external step.