Moshi is a speech-text foundation model and full-duplex spoken dialogue framework.
Run and fine-tune open-source models. Deploy custom models at scale. All with one line of code.
