Published on 2025-06-16
Really impressive model. Another TTS model you can run on your own hardware. The best feature is the instant voice cloning. Previously a lot of other top models like Kokoro required fine tuning the whole model or training a voice with hours of content. With this model you just need a short recording of someone speaking to get a spot on impression of them.
I’ve been testing the model out quite a bit in an audiobook project I’m working on. It’s just a collection of scripts to take any book in any format, clean it up, and turn it into a nice audio book, done end to end with AI. Previously I’ve been using the Kokoro model–since it seemed to be the best freely available–but I’ve now switched over to Chatterbox. Now you can have your favorite audiobook narrator read anything you want. This is going to be game changing for the audio book industry, and I don’t think most people have really understood what’s coming yet.