compare
F5-TTS vs Piper
The same facts for both, read from GitHub every night, and the relation a person reviewed.
| Fact | F5-TTS | Piper |
|---|---|---|
| Language | Python | C++ |
| Licence | MIT | GPL-3.0 |
| Stars | 15k | 5.8k |
| Latest | 1.1.22 | v1.8.0 |
| Last push | 2026-09-21 | 2026-09-28 |
| Release cadence | about 20 days between releases | about 13 days between releases |
| Active contributors | 3 commit authors on the default branch in the last 90 days | 9 commit authors on the default branch in the last 90 days |
| Flags | none | none |
How they relate
Both replace ElevenLabs. Alternatives to ElevenLabs →
PartialF5-TTSVoice cloning from a short reference clip, with a Gradio app.
PartialPiperFast offline voices that run on a Raspberry Pi, without voice cloning.
F5-TTS
- 1.1.222026-07-23fix: allocate fix_duration across chunks to avoid N× duration blow-up
- 1.1.212026-07-05update gradio>=6.15.0 in pyproject.toml
- 1.1.202026-04-20fix: refactor cache handling in DiT, MMDiT, and UNetT classes (lazyinit to avoid EMA deepcopy failure while training)
- 1.1.192026-04-16reuse resamplers and cache vocos MelSpectrogram instances
- 1.1.182026-03-24Add Arabic model details to SHARED.md
Piper
- v1.8.02026-09-04Add Thai phonemizer using TLTK in the new th extra
- v1.7.02026-08-15Add Japanese phonemizer using OpenJTalk (pyopenjtalk-plus) in the new ja extra
- v1.6.12026-08-13Run the g2pW model through piper.g2pw_onnx instead of g2pw.api, dropping torch (~750 MB installed) and requests from th…
- v1.6.02026-07-23Add Hebrew phonemizer using Nakdimon
- v1.5.02026-07-17Add libpiper C++ CLI executable ported from the legacy Piper repository, plus a C++ test suite
F5-TTS
✓ signed The latest release, 1.1.22, carries a signature GitHub verified.
Loading the security report
Piper
unsigned The latest release, v1.8.0, carries no signature GitHub could verify.
Loading the security report