Skip to content

AudioLM

A framework for coherent long-form audio generation

Visit official site

What AudioLM is

The framework is described as a Google Research method for high-quality audio generation that keeps long-term consistency.

What you can do with it

  • Map input audio to discrete token sequences
  • Generate a natural continuation from a short prompt
  • Produce coherent piano music continuations

Getting started

  • Visit the project page for setup guidance
  • Install the required dependencies
  • Download the raw audio waveform datasets

What to watch out for

  • Speaker identity and prosody are preserved in speech
  • No symbolic music representation is used in training

Pros & cons

✓ What we like

  • Speech continuations keep speaker identity and prosody
  • Natural continuations follow a short prompt

! What to watch out for

  • Training uses large amounts of raw audio waveforms
  • The framework works in a discrete token space

FAQ

Who developed it?

The material names Google Research as the developer.

Does it need text?

Speech continuations work without text or annotation.

Last reviewed: 2026-11-15

More AI music creation tools

View all →

How we review