Skip to content
California craftMusic gear & services

Sattari Studio lab

Vox Alpha

Record or drop a vocal, let Vox find the key, then pull every note to the scale with as much or as little strength as you like. Add a third or fifth harmony and export WAV.

Runs on your device. No uploads.

Record

Headphones on. Browser voice processing is switched off.

Or drop a vocal

Dry solo vocal · WAV, MP3, M4A · up to 3 min

Alpha: what it does not do yet

  • One clean, solo (monophonic) voice. Backing tracks, reverb-heavy or polyphonic audio confuse the pitch tracker.
  • Pitch detection is YIN-based (no neural model yet) and the shifter is a simple PSOLA: large corrections and harmonies can sound phasey or robotic.
  • Harmony is one diatonic voice (3rd or 5th above), generated from the same take, so it moves exactly with the lead.
  • Up to 3 minutes per take. Processing is offline (not live while you sing); changes re-render in about a second.
  • Key detection reads the notes you sang and can pick the relative major/minor; check it and override if needed.