Clone a voice
Upload your audio.
And configure any advanced settings.
Enter a name for your new voice.
Select an image
Max file size is 10mb. Only use common images formats like png and jpg.
Max size per file is 50mb. Max total size is 200mb. Only use common formats like mp3 or wav.
This will train an instrument/sound based on your uploaded audio.
This will instantly clone your voice. Only works for speaking right now, singing coming soon...
Isolate vocals
Extract vocals from audio (allows you to upload music with vocals + instruments). Disable if you're uploading raw vocals and hear artifacts in the generated output. Recommended otherwise.
Smart training
Intelligently stops training when the model starts to lose quality. Disable if you want more fine-grained control with "Max minutes" and "Max epochs" below. Recommended otherwise.
Increase this when you want to allow longer training if needed.
60
Increase this when you want to allow more epochs if needed.
1000
Enter the link to the .pth file, or a compressed file (.zip, .rar, etc.) containing the .pth and .index files. We support Hugging Face and Google Drive links at this time. Ask us in the Discord to support more providers.
Select a model
Max file size is 100mb. Only use the .pth file format.
Select an index
Max file size is 200mb. Only use the .index file format.
Optional
By clicking "Train model", you agree to our Terms of Service and Privacy Policy.
View the final output below, along with your previous voices. Click to use.
Upload your vocals and click "Train model" to get started :)
Create
Clone
Stems
Dubs