Moshi is a speech-text foundation model and full-duplex spoken dialogue framework. It uses Mimi, a state-of-the-art streaming neural audio codec.
Repository Index
62,272 visible repositories in the index
1–2 of 2 · Most stars
Sophon-Realtime v1.1: open full-duplex voice models that hear your tone and speak with one. Zero-shot voice cloning, one GPU with vLLM-Omni, OpenAI Realtime / GPT-Live compatible. Apache-2.0.
- full-duplex
- open-source
- paralinguistics
- realtime