Chatterbox is light, quick to clone and easy to steer, which makes it a good everyday choice.
Clone
Each edition has a built-in voice, and any voice you clone can speak with it straight from the recording; no transcript is needed. Turbo wants more than 5 seconds of audio (10 is better); the others do well with about 6. Avocado saves the voice in about a second. Only clone voices you have permission to use.
Expressiveness and pacing
The original and Multilingual editions have two controls: Expressiveness, from flat to dramatic, and Pacing, where lower values are slower and more deliberate.
Turbo tags
Turbo understands nine sounds: laugh, chuckle, sigh, gasp, cough, clear throat, sniff, groan and shush. Pick them from the tag button.
Multilingual
Choose from 23 languages per take. Avocado tells you about each language’s quirks before you generate, for example that Japanese needs kana.
Licence
Chatterbox’s licence is set by Resemble AI. Read it on the Chatterbox model page before you publish anything you make.