Skip to content

GenAU

An audio generation model from Snap Research

Visit official site

What GenAU is

GenAU is described as an audio generation model that combines the AutoCap caption model with the GenAu architecture to improve generated sound.

What you can do with it

  • Generate voices and ambient sounds
  • Try the provided audio examples
  • Tune generation parameters

Model components

  • AutoCap uses audio metadata for captions
  • An audio 1D-VAE produces latent sequences
  • Cross attention and global attention layers are used

Pros & cons

✓ What we like

  • Audio 1D-VAE builds latent sequences from a Mel-Spectrogram
  • Q-Former compresses audio into fewer tokens

! What to watch out for

  • Users start from the official site examples
  • Generation depends on chosen parameters

FAQ

Who developed it?

Snap Research is named as the developer.

What does AutoCap do?

It improves caption quality using audio metadata.

Last reviewed: 2026-11-15

More AI music creation tools

View all →

How we review