Lightning-fast, on-device, multilingual TTS — 32 languages, now including Thai, running natively via ONNX.
Created by Watthanasak Jeamwatthanachai, PhD · EMBA · C|CISO
We added Thai as the 32nd language to Supertone's Supertonic 3 — joining its 31 existing languages without degrading any of them. The samples below showcase seamless Thai ⇄ English code-switching in a single utterance, on-device via ONNX.
This repository is a showcase: listen to the samples below. An interactive demo is coming soon.
Best experience with built-in audio players. The links below also work (each opens GitHub's player on the file page).
Every clip below speaks the same mixed Thai + English line:
สวัสดีครับ ผมชื่อสมชาย และผมชอบเรียนภาษาไทย — A gentle breeze moved through the open window while everyone listened to the story.
| Voice | 🇹🇭 Thai-native (th) |
🇬🇧 English-native (en) |
🌐 Neutral bilingual (na) |
|---|---|---|---|
| F1 (female) | ▶ play | ▶ play | ▶ play |
| M1 (male) | ▶ play | ▶ play | ▶ play |
| F3 (female) | ▶ play | ▶ play | ▶ play |
For inline players, open the showcase page.
The same model speaks the same text three ways — pick whichever sounds most natural for your use case:
th— Thai-native speaker. Thai is flawless; English carries a light Thai accent.en— English-native bilingual. English is flawless; Thai carries a light English accent.na— neutral bilingual. Cleanest "professional bilingual" delivery across both languages.
- ✅ 32 languages — all 31 of Supertonic 3's original languages preserved, with Thai added as the 32nd.
- ✅ Adding Thai degraded none of the original languages — they remain at native quality.
- ✅ Thai + English in a single utterance — natural code-switching, no stitching.
- ✅ Multiple voices — male and female, with voice identity preserved across languages.
- ✅ On-device — runs natively via ONNX, no cloud required.
- ✅ 44.1 kHz output.
Coming soon — a live, in-browser demo where you can type your own Thai/English text and hear it. (This section will be updated once the demo is hosted.)
Created by Watthanasak Jeamwatthanachai, PhD · EMBA · C|CISO.
Built on top of Supertonic 3 by Supertone Inc. The base model is released under OpenRAIL-M. The audio in this repository consists of outputs from our Thai-adapted model and is provided for demonstration under the terms in LICENSE.