Chatterbox Turbo is an open-source voice model built by Resemble AI. It's meant to be fast and ready to use in production with built-in safety like watermarking and support for over 23 languages. It runs under the MIT license and gives devs more control and a clear look at how it works.
Tech specs.
It does text-to-speech almost 6× faster than real time on GPU with just 75ms delay.It can copy a voice using only about 5 seconds of audio, no extra training needed.
Paralinguistic tags.
This is the standout part. Other tools don’t really do this. You can change how emotional the voice sounds with one setting, from flat to dramatic. It lets you type tags like [gasp], [laugh], or [cough] to make the voice react naturally.
Usage. Good for anything that needs fast or live voice output, like:
If you'd like to access this model, you can explore the following possibilities:
Spaces