How it works

Four steps, one history.

Describe it, check the cost, listen, keep.

Four steps

The four steps.

  1. Describe

    A voice-over, a music bed, an effect or a dialogue script.

  2. See the cost

    The credits are shown before anything is made.

  3. Listen

    Takes appear one by one with waveforms.

  4. Keep

    Collections, downloads, and Geenius OS Media with provenance.

Where the work happens.

AI Audio is where the composer, the takes and the library live. A take is made by a provider: today the simulated Preview in your browser, and a real model where one is switched on. Your own AI can queue runs through Geenius OS.

  • AI Audio: composer, history, library
  • The provider: Preview, or a model
  • You: describe and keep
your AI · audioqueued
a browser Preview run waits for an open page

Try the loop yourself.