GPT-SoVITS
github.com
- Free plan
- Windows
- macOS
- Linux
- Web
- API
- Self-hosted
What GPT-SoVITS is
GPT-SoVITS is a free WebUI for voice conversion and text-to-speech. It can turn a five-second vocal sample into speech, or use one minute of training data to improve voice similarity and realism. Cross-language inference supports English, Japanese, Korean, Cantonese, and Chinese. The WebUI includes tools for separating accompaniment, segmenting training sets, multilingual speech recognition, and text labeling. Speech-recognition options include Fun-ASR-Nano, SenseVoice, classic FunASR, and Faster Whisper. The repository also provides an API with GET and POST inference endpoints that return WAV audio streams on success. Installation instructions cover Windows, Linux, and macOS, and Docker Compose configurations are documented for full and Lite services. The Lite Docker image omits ASR and UVR5 models; UVR5 models must be downloaded manually, while ASR models download as needed. On macOS, the project temporarily uses CPUs because it reports lower quality from models trained with Mac GPUs. GPT-SoVITS is released under the MIT License, which allows use, modification, distribution, sublicensing, and sale subject to its conditions. The README links to Chinese and English user guides.
Who it suits
GPT-SoVITS may suit people building voice datasets or text-to-speech workflows, including beginners using its integrated WebUI tools. It also offers an API and deployment options for users working with existing systems.
Good to know, and what to watch
- Five-second sample supports instant text-to-speech conversion.
- Cross-language inference covers five listed languages.
- WebUI includes dataset preparation and speech-recognition tools.
- MIT License permits commercial use under its conditions.
- Lite Docker image excludes ASR and UVR5 models.
- UVR5 models require manual download for Lite.
- macOS temporarily uses CPUs due to reported GPU-model quality.
The verdict
GPT-SoVITS brings voice conversion, speech generation, and dataset tools into a free WebUI with API options. Users choosing Lite deployment or macOS should account for the stated model and processing limitations.
What the maker publishes
- API
- The repository includes an API exposing GET and POST inference endpoints that return WAV audio streams on success. source
- API access
- Yes source
- ASR integrations
- The WebUI offers Fun-ASR-Nano, SenseVoice and classic FunASR backends, and Faster Whisper is also available. source
- ASR integrations
- The WebUI lists Fun-ASR-Nano, SenseVoice, and classic FunASR; Faster Whisper is also available as an ASR backend. source
- Commercial use
- Yes source
- Dataset tools
- WebUI tools include accompaniment separation, automatic training-set segmentation, multilingual ASR, and text labeling. source
- Deployment
- The project documents Docker images and Docker Compose services, including full and Lite variants. source
- Docker
- Docker Compose defines full and Lite services for CUDA 12.6 and CUDA 12.8 environments. source
- Export formats
- wav,ogg,aac source
- Few-shot TTS
- The model can be fine-tuned with 1 minute of training data for improved voice similarity and realism. source
- Few-shot TTS
- The project says users can fine-tune with one minute of training data to improve voice similarity and realism. source
- Free plan
- Yes source
- Intended users
- The integrated WebUI tools are described as assisting beginners in creating training datasets and GPT/SoVITS models. source
- Languages
- Cross-lingual inference supports English, Japanese, Korean, Cantonese and Chinese. source
- Languages
- Cross-lingual inference supports English, Japanese, Korean, Cantonese, and Chinese. source
- License
- The software is released under the MIT License, which permits use, copying, modification, distribution, sublicensing and sale subject to the license conditions. source
- License
- The repository identifies its license as MIT. source
- Lite limit
- The Lite Docker image does not include ASR or UVR5 models; UVR5 models must be downloaded manually and ASR models download as needed. source
- Local API
- The repository includes api.py and api_v2.py alongside the WebUI. source
- macOS limit
- The README says models trained with Mac GPUs produce significantly lower quality than models trained on other devices, so it temporarily uses CPUs instead. source
- macOS limitation
- The project says models trained with GPUs on Macs have significantly lower quality and therefore temporarily uses CPUs on macOS. source
- Operating systems
- Installation instructions are provided for Windows, Linux, and macOS. source
- Platforms
- web,windows,macos,linux,api source
- Product
- GPT-SoVITS is a WebUI for few-shot voice conversion and text-to-speech. source
Checked against the maker's pages on 2 Oct 2026.