so-vits-svc
Open-source SoftVC VITS singing voice conversion, run fully offline.
Retired
so-vits-svc appears to be discontinued or shut down — archived by maintainers; community forks continue This page is kept as an archive record and is excluded from search engines.
About
so-vits-svc (SoftVC VITS Singing Voice Conversion) is an open-source project that converts one singing voice into another, distinct from text-to-speech VITS. It runs completely offline and became one of the most widely used singing voice conversion frameworks, with a large community producing models and forks.
The project was officially archived by its maintainers, but community forks continue development, including versions with improved interfaces and real-time conversion clients. Note: the project is AGPL-3.0 licensed and intended by its developers for fictional characters; cloning real voices without consent raises legal and ethical issues.
What it can do
- Singing voice conversion between voices
- Fully offline inference
- Community forks with improved UIs
- Real-time conversion clients available
Setup
Clone the repo or a maintained fork, install dependencies, and run training or inference scripts.
Prerequisites
Python, PyTorch, and a GPU recommended; training data for the target voice.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 103; Trending score 70/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.