supertone-inc/supertonic
An on-device text-to-speech library that runs ONNX models offline, with a 99M-parameter model and 31 languages. Development ended in July 2026, and the repo is being archived, so it will not get further updates.
Idea adaptedKey files read, not run
Tested in Vietnamese on 31 July 2026 against edge-tts in the same run: 22.0% word error rate against 9.9%, and not faster (RTF 0.16 against 0.14). 15 of the 40.4 s of audio was silence. Development had already ended on 23 July 2026, before the test.
How we know
We read the README and key files, not every source file, and did not run it, so the most we can say is Idea adapted or Risky.
Good for
Anyone who already runs it and only needs offline speech in an app that is already stable.
Not for
New projects. The maintainer has ended development, and the Voice Builder for custom voices was scheduled to close on 31 August 2026.
Repo
- Stars
- 13,775
- License
- MITUse it, change it, sell it, closed-source included. Keep the author's copyright notice.
- Maturity
- archived
- Last push
- 2026-09-09
- Last checked
- 2026-11-14