Pocket TTS: A high quality TTS that gives your CPU a voice(kyutai.org)
594 points by pain_perdu | 142 comments
Kyutai blog post introducing Pocket TTS, a 100M-parameter text-to-speech model with voice cloning that runs in real time on a laptop CPU.
#kyutai on Hacker News
3 stories tagged kyutai, newest first.
Neural audio codecs: how to get audio into LLMs(kyutai.org)
414 points by karimf | 115 comments
Kyutai explainer article by Václav Volhejn on how neural audio codecs turn audio into tokens so it can be modeled by large language models.
Moshi: A speech-text foundation model for real time dialogue(github.com)
331 points by gkucsko | 56 comments
GitHub repository for Kyutai Labs' Moshi, a speech-text foundation model for real-time spoken dialogue, with Python, MLX, Rust, and client code.