Skip to content

Bark

A transformer text-to-audio model with many voices

Visit official site

What Bark is

The model is described as a transformer-based text-to-audio model from Suno that produces realistic multilingual speech plus other audio such as music and sound effects.

What you can do with it

  • Generate speech in several languages
  • Produce music, noise and simple effects
  • Add laughter, sighs or crying sounds

Getting started

  • Install the needed libraries and the model
  • Call preload_models() to fetch the models
  • Use generate_audio() on a text prompt
  • Save the result with write_wav()

Pros & cons

✓ What we like

  • Long audio generation is supported
  • Pretrained checkpoints are available for commercial use

! What to watch out for

  • CPU and GPU runs have different hardware requirements
  • Voice choice comes from the preset list

FAQ

What can it generate?

The material names multilingual speech, music, background noise, simple effects, laughter, sighs and crying.

How is audio produced?

The steps name preload_models(), generate_audio() and write_wav().

Last reviewed: 2026-11-15

More AI audio tools tools

View all →

How we review