Supertone

/

supertonic-3

like 199

Follow

Supertone 378

Hi @devops724 ,

Thanks for your interest.

At the moment, Persian/Farsi is not supported by Supertonic 3. The current release supports 31 languages, but fa is not included yet.

We also do not currently provide training or fine-tuning code, example datasets, or a public fine-tuning guide. The open-weight release is focused on ONNX inference with the provided model files and preset voice styles.

That said, we are interested in improving multilingual coverage over time. If you know of high-quality open Persian speech datasets, feel free to share them and we can review them internally for future updates.

Best regards,
Supertone team

i will be glad to help in this case

which format do you prefer for dataset?
is there any guide about structure , audio quality ... let me know what i should follow to provide dataset
this help to reduce preprocess efforts in your side

Hi @devops724 ,

Thank you, we really appreciate the offer.

For our internal data pipeline, native Persian/Farsi speech audio itself would already be useful. If needed, we can handle filtering, ASR, and other processing internally to turn raw audio into a usable training dataset.

Optional metadata such as text transcripts, speaker labels, or license information would of course be helpful, but they are not strictly required for us to take an initial look.

So you do not need to prepare the data in a specific format for us. If you know of Persian/Farsi speech data that is high quality and can be shared, feel free to send links or details and we can review them internally.

Just to set expectations clearly: sharing data does not mean we will immediately train or release Persian/Farsi support, but it would help us evaluate the possibility for future multilingual updates.

Thanks again for your willingness to help.