Kyutai's Speech-To-Text and Text-To-Speech models based on the Delayed Streams Modeling framework.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).