Kitten TTS

AI Tools

25MB CPU-only open-source TTS model with 15M parameters and 8 voices

Kitten TTS is an ultra-lightweight open-source text-to-speech model with only 15M parameters and 25MB size, running on pure CPU without GPU. Supports 8 different voices with fast inference. Ideal for embedded devices and edge deployment. Compared to multi-GB TTS models, Kitten TTS compresses to the extreme while maintaining usable audio quality. 1003 points on HN.

← Back to directory

Kitten TTS

AI Tools

25MB CPU-only open-source TTS model with 15M parameters and 8 voices

https://github.com/KittenML/KittenTTS

Kitten TTS is an ultra-lightweight open-source text-to-speech model with only 15M parameters and 25MB size, running on pure CPU without GPU. Supports 8 different voices with fast inference. Ideal for embedded devices and edge deployment. Compared to multi-GB TTS models, Kitten TTS compresses to the extreme while maintaining usable audio quality. 1003 points on HN.