GPT-SoVITS

About GPT-SoVITS

GPT-SoVITS combines GPT-style language modeling with SoVITS voice conversion to achieve state-of-the-art few-shot voice cloning. With just 3-10 seconds of reference audio plus a transcript, it produces remarkably natural speech that closely matches the target voice. It excels at cross-lingual synthesis - train on one language and generate in another.

Key Features

Use Cases

How to Use GPT-SoVITS

1. 1

   Create a free account or open the demo

   Sign up for TextToSpeechAI to receive free starter credits, or jump straight into the demo to try GPT-SoVITS with no signup required.

2. 2

   Select GPT-SoVITS and upload a reference clip

   Choose GPT-SoVITS as your engine, then upload a 3-10 second reference clip of the voice you want to clone. Adding the transcript of that clip gives the cleanest, most accurate clone.

3. 3

   Enter your text

   Type or paste the text you want spoken in the cloned voice. GPT-SoVITS supports Chinese, English, Japanese, Korean, and Cantonese, including cross-lingual cloning from a reference in another language.

4. 4

   Generate the audio

   Click generate to send the job to our GPU servers. GPT-SoVITS renders excellent-quality cloned speech at medium speed, with 25 credits billed per 1,000 characters.

5. 5

   Download or use the API

   Download your finished GPT-SoVITS audio as a file, or automate generation through the TextToSpeechAI REST API at api.texttospeechai.com for production workflows.

GPT-SoVITS API

Generate speech programmatically using the TextToSpeechAI REST API.

curl -X POST "https://api.texttospeechai.com/v1/generate/" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "GPT\u002DSoVITS produces the highest quality voice cloning from just a few seconds of audio.",
    "voice": "en_US-lessac-medium"
  }'

Frequently Asked Questions