Overview
Bark is a text-to-audio model released by Suno in April 2023. Unlike conventional text-to-speech systems, it generates audio directly from text with a GPT-style architecture, which lets it produce laughter, sighs, simple music and background sounds as well as speech in multiple languages. It is available under the MIT license.
Highlights
- Multilingual speech generation
- Non-verbal sounds such as laughing and sighing
- Simple music and sound effects
- Speaker presets for different voices
- MIT license permitting commercial use
Typical uses
Experimental voice and sound generation, creative projects and research on generative audio.
Things to consider
- Outputs can be unpredictable and may deviate from the script.
- Newer text-to-speech models offer more control and quality.
- Use synthetic voices responsibly and with disclosure.
Specifications are taken from the provider’s official announcement, documentation or model card. Benchmark results depend on evaluation settings; see how AI benchmarks work.