CosyVoice

What CosyVoice is

CosyVoice is a multilingual text-to-speech system for inference, training, and deployment. Its Fun-CosyVoice 3.0 version is designed for zero-shot multilingual speech synthesis across nine languages and more than 18 Chinese dialects or accents, with multilingual and cross-lingual voice cloning. Users can guide language, dialect, emotion, speed, and volume, and adjust pronunciation with Chinese Pinyin or English CMU phonemes. It handles numbers, special symbols, and varied text formats without a traditional frontend module. Text-in, audio-out streaming has stated latency as low as 150 ms. The repository documents Conda installation with Python 3.10, pretrained model downloads, and Docker deployment using gRPC or FastAPI, as well as NVIDIA Triton and TensorRT-LLM acceleration. It is Apache License 2.0 software with self-managed installation and deployment; no paid plans are listed. The maintainers direct users to GitHub Issues and an official Dingding chat group. The repository says its content is for academic purposes and to demonstrate technical capabilities.