Built by a producer, not a committee.

VoxBrain started from a frustration every producer knows: the vocal is the record, and getting it right takes an hour of the same moves you made yesterday. Cut the mud. Tame the esses. Get the level under control. Do it again tomorrow on a different voice, where all the numbers are different.

The tools that claim to automate that mostly do not. They apply a preset and call it intelligence — the same chain whatever you feed them, with no idea whether your take is noisy, dull, sibilant or already fine. Worse, they will not tell you what they did, so you learn nothing and you cannot argue.

What we believe

Measure first, decide second. Every move comes from a measurement of the recording in front of it. If your noise floor is −60 dB you do not get a gate. If it is −30 dB you get one, set to your floor.

Show your working. The report opens with what the plugin heard before what it did about it. A conclusion without evidence is a preset wearing a lab coat.

The producer is in charge. Lock anything and the AI will not touch it — and it will say so. Nothing is hidden, nothing is irreversible, and one Undo takes back a whole pass.

You own what you bought. Perpetual licences. No subscription, no phone-home, no feature that stops working because a server went away.