aoiandroid/Supertone-supertonic-3-mirror

Supertonic 3 | Lightning Fast, On-Device, Accurate TTS

[image]

<p align="center"> <a href="https://huggingface.co/spaces/Supertone/supertonic-3"><img src="https://img.shields.io/badge/Demo-Hugging_Face-yellow?style=for-the-badge" alt="Demo"></a> <a href="https://github.com/supertone-inc/supertonic"><img src="https://img.shields.io/badge/Code-GitHub-black?style=for-the-badge&logo=github" alt="Code"></a> <a href="https://pypi.org/project/supertonic/"><img src="https://img.shields.io/badge/Python-SDK-blue?style=for-the-badge&logo=python" alt="Python SDK"></a> </p>

Supertonic is a lightweight text-to-speech system for local inference. It runs with ONNX Runtime entirely on your device, with no cloud call required for synthesis.

Supertonic 3 expands the open-weight release from 5 to 31 languages, improves reading stability, and reduces repeat/skip failures.

Quick Start

Install the Python SDK and generate speech immediately. On first run, the SDK downloads the model assets from Hugging Face.

pip install supertonic

from supertonic import TTS

tts = TTS(auto_download=True)
style = tts.get_voice_style(voice_name="M1")

text = "A gentle breeze moved through the open window while everyone listened to the story."
wav, duration = tts.synthesize(text, voice_style=style, lang="en")

tts.save_audio(wav, "output.wav")
print(f"Generated {duration:.2f}s of audio")

What's New in Supertonic 3

• —31 languages: expanded from the 5-language Supertonic 2 release.
• —More stable reading: fewer repeat and skip failures, especially on short and long utterances.
• —Higher speaker similarity: improved similarity across the shared-language set compared with Supertonic 2.
• —Expression tags: supports simple tags such as <laugh>, <breath>, and <sigh>.

Performance Highlights

Supertonic 3 is designed for practical on-device inference: compact enough to run locally, while staying competitive with much larger open TTS systems.

Reading Accuracy

<p align="center"> <img src="img/metrics/s3vsmeasuredwerrange_voxcpm2.png" alt="Supertonic 3 reading accuracy compared with measured model ranges and VoxCPM2"> </p>

Across measured languages, Supertonic 3 stays within a competitive WER/CER range against much larger open TTS models such as VoxCPM2, while preserving a lightweight on-device deployment path. Asterisked languages use CER; the others use WER.

Supertonic 2 to Supertonic 3

<p align="center"> <img src="img/metrics/supertonic2vs3_comparison.png" alt="Supertonic 2 and Supertonic 3 comparison"> </p>

Compared with Supertonic 2, Supertonic 3 reduces repeat and skip failures, improves speaker similarity across the shared-language set, and expands language coverage from 5 to 31 languages.

Runtime Footprint

<p align="center"> <img src="img/metrics/runtimecpugpulatencymemory.png" alt="Supertonic CPU runtime compared with GPU baselines"> </p>

Supertonic 3 runs fast on CPU, even compared with larger baselines measured on A100 GPU, and uses substantially less memory. It does not require a GPU, which makes local, browser, and edge deployment much easier.

Model Size

<p align="center"> <img src="img/metrics/modelsizecomparison.png" alt="Model size comparison"> </p>

At about 99M parameters across the public ONNX assets, Supertonic 3 is much smaller than 0.7B to 2B class open TTS systems. The smaller model size is a practical advantage for download size, startup time, and on-device inference.

Supported Languages

License

This project's sample code is released under the MIT License. See the GitHub repository for details.

The accompanying model is released under the OpenRAIL-M License. See the LICENSE file in this repository for details.

This model was trained using PyTorch, which is licensed under the BSD 3-Clause License but is not redistributed with this project. See the PyTorch license for details.

Copyright (c) 2026 Supertone Inc.