GUIDE

A GPT-SoVITS alternative for 31 languages

GPT-SoVITS is one of the most popular open-source voice cloning projects: a 5-second sample for instant cloning, about a minute of speech for fine-tuning and a WebUI with tools to prepare training data. It also covers five languages, and a good voice from it usually means preparing data on a computer with a graphics card.

The free SuperClonica engine runs the Supertonic 3 model in 31 languages on an ordinary processor, with a catalog of ready voices. This guide compares the two honestly and says when GPT-SoVITS is the better pick.

Browse voices Install the SuperClonica engine

GPT-SoVITS in short

It is an excellent tool for Chinese and Japanese voices and for people who enjoy preparing datasets.

  • Zero-shot cloning from a 5-second sample, and fine-tuning on about one minute of speech.
  • Five languages: Chinese, English, Japanese, Korean and Cantonese.
  • A WebUI with tools to separate vocals, slice recordings and label training data.
  • MIT license, actively developed; fast on NVIDIA graphics cards, slower on processors.

Compare the sound free

On the page of any voice, paste a line you voiced with GPT-SoVITS, 30 to 100 characters, choose the language and press Voice it. In a few seconds you can compare, with no account.

What SuperClonica does differently

Voicing runs on an ordinary processor: in our tests a short phrase took 0.43 s on four threads of a Ryzen 7 5800X.

  • 31 languages, including Spanish, German, French, Russian, Portuguese, Italian, Polish and Hindi; Chinese and Cantonese are not among them.
  • Ready voices: ten stock voices and more than 170 catalog voices with free demos, no dataset needed.
  • Each voice is a JSON file of about 0.3 MB and speaks every language of the model.
  • One command to install: python -m pip install superclonica-cpu, or a Docker image with an HTTP API.

Your own voice

GPT-SoVITS imitates a voice from a sample at generation time. In SuperClonica, a voice is trained once in the studio from a clean 5 to 10 second recording, on a computer with an NVIDIA graphics card, or ordered on the exchange from an author who trains it. After that it runs on any processor and in all 31 languages, with a trace of the accent of the original recording. Clone only voices whose owners agreed.

Which one to choose

Choose GPT-SoVITS for Chinese, Cantonese and Japanese voices, for instant cloning from a sample, and when you have an NVIDIA graphics card. Choose SuperClonica for European and the other supported languages, for ready character and narrator voices, and for simple deployment on ordinary servers. Nothing stops you from using both: one for Chinese and Cantonese, the other for the rest.