Skip to content
Get packs

Supertonic

Open-source, on-device text-to-speech that runs locally via ONNX; archived and no longer maintained.

GitHub Download v2.0.0
Stars
13.8k
Forks
1.6k
Licence
MIT
Language
Swift
Last update
2 months ago
Type
App

This repository is archived and no longer maintained.

What it is

Supertonic is a text-to-speech system that turns written text into spoken audio entirely on your own device, using ONNX Runtime, so no cloud service or API call is needed. The README describes a 99M-parameter model supporting 31 languages and outputting 44.1kHz WAV audio. This repository is archived, so it is kept for reference and gets no updates, bug fixes, or support.

  • Generate a WAV file of a sentence locally using the Python example in the py folder
  • Compare speech output across the 31 supported languages
  • Run the model in a Python script to batch-synthesize text

Great for

  • Developers who want to experiment with local, offline speech synthesis
  • People studying how an ONNX-based TTS pipeline is put together
  • Projects that need a fixed, pinned snapshot of the archived Supertonic 3 model
  • Multilingual prototypes that can accept the archived model's limits

Not ideal for

  • Production systems needing security patches or bug fixes, since the archive states none will be provided
  • Users who need active support, since issues and pull requests are no longer monitored
  • Anyone wanting hosted demos or Voice Builder services, which the archive does not include

How to get it

  1. 1

    Clone and build from source

    You'll need: Python 3.11 · Git

    Clone the source

    git clone https://github.com/supertone-oss-archive/supertonic.git

    Install the download tool in a virtual environment

    python -m pip install huggingface_hub

    Download the archived Supertonic 3 model

    hf download supertone-oss-archive/supertonic-3 --revision aafc6e32416a594460b32413efc49d7fe4ce6d46 --local-dir assets

    Install Python example dependencies

    python -m pip install -r py/requirements.txt
  2. 2

    Run it

    • Clone the repository and create a Python 3.11 virtual environment
    • Install huggingface_hub and download the pinned archived model into assets/
    • Install the requirements in py/requirements.txt
    • From the py folder, run example_onnx.py with --text and --lang en
    • Find the generated WAV in py/results/
  3. 3

    Get the most out of it

    • Use the --n-test and --text options of example_onnx.py to control output
    • Read py/README.md for voice selection and batch synthesis options
    • Use lang="na" when the input language is unknown, per the README
    • Try the optional Python SDK with auto_download=False and a local model_dir, as described in the README

Downloads

Source code (.zip)main branch · straight from GitHubRelease v2.0.0Release notes and sourceOfficial website & docshuggingface.co/spaces/Supertone/supertonic-3

Strengths

  • Runs locally with no network connection needed after models are downloaded, per the README
  • Ships examples across many languages and runtimes, including Python, Node.js, Go, Rust, and Swift
  • MIT-licensed code, with model weights under their own stated licenses

Trade-offs

  • Archived: no updates, security patches, or support from the original authors
  • Model downloads depend on the archive's Hugging Face namespace, and the README notes the original hosted services are no longer provided

How it compares

ProjectStarsTypeLevelUpdatedBest for
supertonic (this)13.8kAppintermediate2 months agoDevelopers who want to experiment with local, offline speech synthesis
calculator31.1kAppintermediate14 days agoWindows 11 users who want to read or modify the built-in calculator
CodeEdit23.1kAppbeginner20 days agoDevelopers on a Mac who work on non-Apple projects and want a native macOS editor
AudioKit11.5kLibraryintermediate3 months agoSwift developers adding synthesizers, effects, or audio analysis to an iOS or macOS app
GPT-SoVITS64.6kAppintermediate6 days agoHobbyists who want to experiment with cloning a voice they have permission to use
NativeScript25.7kFrameworkintermediate5 days agoJavaScript or TypeScript developers who want to build native iOS or Android apps

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.