myshell-ai/OpenVoice
OpenVoice clones the tone color of a short reference clip and synthesises speech with control over emotion, accent, rhythm, pauses and intonation. Version 2 improves audio quality and natively supports English, Spanish, French, Chinese, Japanese and Korean, including cross-lingual cloning where neither language appeared in training. The code, weights and usage guides are published under the MIT license for research and commercial use.
Maintenance activity