Bark

Bark is a transformer-based text-to-audio model created by Suno. Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying. To support the research community, we are providing access to pretrained model checkpoints, which are ready for inference and available for commercial use.

It may deviate from prompts in unexpected ways. It is licensed under the MIT License.

Bark aims to be used for research and demo purposes, offering access to model checkpoints for inference. It supports multiple languages, voice presets, and longer audio generation, and it notes usage caveats and safety considerations. The project provides installation instructions, usage examples, and references to community resources such as a Discord channel.

Bark is a GPT-style model handling audio generation from prompts, with hardware notes, language capabilities, and a note about the model’s variability compared with traditional TTS. It includes a FAQ and links to notebooks, demos, and a playground in development.

Use cases

• Text-to-audio generation for research and demos