#5353rd of 121 in VoiceFor voice, speech and music
vieneu-tts
Vietnamese text-to-speech with voice cloning and an OpenAI-compatible streaming API
- Stars
- 3.5k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- Python
1,322 on 29 Apr 20263,529 on 18 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
VieNeu-TTS is a Vietnamese (and bilingual Vietnamese-English) TTS model with a Python SDK, a Gradio web UI and an API server exposing POST /v1/audio/speech. v3 Turbo outputs 48 kHz audio, ships 25 preset voices, clones a voice from a 3–8 s clip, and runs on CPU via ONNX Runtime or on CUDA via PyTorch. LoRA fine-tuning is supported; the newer v4 is proprietary and only available through vieneu.io.
Who it is for: Developers needing self-hosted Vietnamese TTS with cloning and streaming
Strengths
- Torch-free CPU install via ONNX Runtime; RTF about 0.5 on a 12th-gen Core i5
- OpenAI-compatible /v1/audio/speech with streaming; first audio about 115 ms on RTX 3060
- Voice cloning from a 3–8 s clip with automatic denoising, plus 25 preset voices
- LoRA fine-tuning on a single consumer GPU; v3 Nano option for weak CPUs
Weaknesses
- Focused on Vietnamese; the README describes no other languages beyond English
- Newer v4 with closer cloning fidelity is proprietary and API-only
- Dubbing, Lecture, Tiktok Reel and Audiobook features are only in the official app
- int8 CPU mode needs VNNI support and can produce garbled audio on older CPUs
What it needs
- GPU optional
- Docker + Compose
- Needs ONNX Runtime, PyTorch (GPU path), uv
- Models: VieNeu-TTS v3 Turbo, VieNeu-TTS v3 Nano, VieNeu-TTS v2, VieNeu-TTS v1
- port 7860
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Known problems
Open now, reported in the last 90 days
- Bị lặp âm thanh ↗ (opens in a new tab)1 reactions
- v3 Turbo: "chánh" đọc thành "tránh" (sea_g2p gộp ch/tr thành cùng âm tʃ; phoneme_dict không có tác dụng) / "chánh" rendered as "tránh" ↗ (opens in a new tab)1 reactions
- [Thank you] create vi- full-duplex benchmark using the technology ↗ (opens in a new tab)1 reactions
Titles are the reporters' own words, taken from the project's GitHub issues and ranked by reactions. We have not tried to reproduce any of them.
Also in Voice
See all 121| Rank | Project | Score |
|---|---|---|
| 1 | voicestudioDesktop app for voice cloning, dubbing, dictation and audiobooks | 98 out of 100 |
| 2 | handyOffline push-to-talk dictation app that pastes transcripts into any text field | 93 out of 100 |
| 3 | whisper.cppC/C++ speech-to-text engine running OpenAI Whisper models locally | 91 out of 100 |
| #4 | meetilyDesktop app that records, transcribes and summarizes meetings locally | 88 out of 100 |
| #5 | agentsPython framework for realtime voice and multimodal agents on LiveKit WebRTC | 86 out of 100 |