Supertone/supertonic-3 · Hugging Face

Source: https://huggingface.co/Supertone/supertonic-3

https://huggingface.co/Supertone/supertonic-3#supertonic-3–lightning-fast-on-device-accurate-ttsSupertonic 3 | Lightning Fast, On-Device, Accurate TTS

Supertonicis 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 3expands the open-weight release from 5 to31 languages, improves reading stability, and reduces repeat/skip failures.

https://huggingface.co/Supertone/supertonic-3#quick-startQuick 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")

https://huggingface.co/Supertone/supertonic-3#whats-new-in-supertonic-3What’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\>.

https://huggingface.co/Supertone/supertonic-3#performance-highlightsPerformance Highlights

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

https://huggingface.co/Supertone/supertonic-3#reading-accuracyReading Accuracy

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.

https://huggingface.co/Supertone/supertonic-3#supertonic-2-to-supertonic-3Supertonic 2 to Supertonic 3

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.

https://huggingface.co/Supertone/supertonic-3#runtime-footprintRuntime Footprint

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.

https://huggingface.co/Supertone/supertonic-3#model-sizeModel Size

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.

https://huggingface.co/Supertone/supertonic-3#supported-languagesSupported Languages

CodeLanguageCodeLanguageCodeLanguageCodeLanguageenEnglishkoKoreanjaJapanesearArabicbgBulgariancsCzechdaDanishdeGermanelGreekesSpanishetEstonianfiFinnishfrFrenchhiHindihrCroatianhuHungarianidIndonesianitItalianltLithuanianlvLatviannlDutchplPolishptPortugueseroRomanianruRussianskSlovakslSloveniansvSwedishtrTurkishukUkrainianviVietnamese

https://huggingface.co/Supertone/supertonic-3#licenseLicense

This project’s sample code is released under the MIT License. See theGitHub repositoryfor details.

The accompanying model is released under the OpenRAIL-M License. See theLICENSEfile 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 thePyTorch licensefor details.

Copyright (c) 2026 Supertone Inc.