Between June 2025 and February 2026, we ran the Unmute Voice Donation Project: an initiative to collect voice recordings for our open-source text-to-speech (TTS) models. We are grateful to everyone who donated their voices and contributed to this project. The voice donation period has now ended, and we are no longer accepting new voice recordings. In total, there were 374 voices submitted, out of which 228 passed verification. A big thank you to all participants ❤️
The voices are available in the kyutai/tts-voices Hugging Face repo, specifically the voice-donations/ directory. You can also try out how they sound with our text-to-speech models here.
The project was run as part of the release of Kyutai TTS 1.6B, a model for which we did not publicly release voice cloning support. Instead, we encouraged people to contribute to the Unmute Voice Donation Project, and provided the voices for anyone to use with the Kyutai TTS 1.6B model.
In case you need to revisit them, here are the Terms of Use and Privacy Policy of the Project, as well as the archived version of the donation page.
If you use the voice donations in academic work, please cite:
@misc{unmute_voice_donation,
title={Unmute Voice Donation Project},
author={Volhejn, V\'{a}clav},
year={2025},
url={https://unmute.sh/voice-donation}
}