Audio
Real-Time-Voice-Cloning
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
★ 60.2k starsNOASSERTION
View guide →Vietnamese TTS with instant voice cloning — on-device, real-time CPU inference, 48 kHz audio
VieNeu-TTS is an open-source audio project by pnnbao97 (pnnbao97/VieNeu-TTS on GitHub). It has 2,787 GitHub stars and 862 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 30 September 2026.
Choose a built-in voice, plan to clone a voice from a short clip, or run the streaming API server for an app.
Install the vieneu SDK from PyPI for code use, or clone the repo and sync dependencies with uv for the Web UI.
Pick the CPU path (uv sync / pip install vieneu) or the GPU path (uv sync --extra cuda); Windows users can also use the standalone installer from vieneu.io.
Call Vieneu().infer(...) with Vietnamese text and a voice name, or start the Web UI and generate from the browser.
The current open-source architecture (with its own audio tokenizer and phonemizer) outputs 48 kHz audio, up from 24 kHz in v2.
Built-in voices cover Northern, Central and Southern accents, with 10 editors' picks recommended as starting points — no reference clip needed.
Clone a voice from a 3–8 second clip, with automatic reference denoising.
infer_stream streams frame by frame on CPU and GPU, and an OpenAI-compatible POST /v1/audio/speech API works with the OpenAI SDK, Pipecat and LiveKit.
Batched generation (up to batch size 32) including a multi-speaker conversation mode that batches a whole script.
Train your own voice or reading style on one consumer GPU.
Clone VieNeu-TTS for the Web UI / full repo experience; SDK-only users can skip to pip install vieneu.
git clone https://github.com/pnnbao97/VieNeu-TTS.git
cd VieNeu-TTSUse uv as the README recommends — CPU/minimal, or with the cuda extra for GPU.
uv sync
# GPU: uv sync --extra cudaLaunch the Gradio UI, or pip install vieneu and call Vieneu().infer(text, voice=...) from Python.
uv run vieneu-web # http://127.0.0.1:7860
# SDK: pip install vieneuIt is created and maintained by pnnbao97 (Phạm Nguyễn Ngọc Bảo); the commercial v4 and app live at vieneu.io.
The code in this repo (v3 Turbo) is Apache-2.0 licensed and free. The README states v4 is proprietary and only offered through the VieNeu API/site.
No — the open-source version runs on-device (CPU via ONNX Runtime, or your own GPU). No cloud key is required.
gTTS calls Google Translate's hosted speech endpoint with fixed voices. VieNeu is a local Vietnamese-first model with preset regional voices, cloning, streaming and much higher audio quality (48 kHz).
pip install vieneu, then vieneu = Vieneu(); audio = vieneu.infer(text, voice=...) — or clone and run uv run vieneu-web.
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
MOSS-TTS family: open text-to-speech models for voice cloning, dialogue, voice design, and real-time agents
Open-source real-time spoken dialogue with personalized voice cloning.
MOSS-TTS-Nano: a 0.1B open multilingual speech model that clones voices and streams in real time on CPU
A pure C++ ggml inference engine for audio models — TTS, STT, VAD, voice conversion and music.
Fish Speech — Fish Audio's expressive open text-to-speech and voice cloning: SOTA-quality TTS you can self-host, under the Fish Audio Community License (check terms)
The most-starred other Audio projects in the LibreAI catalogue — the realistic shortlist if VieNeu-TTS is not the right fit.
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Robust Speech Recognition via Large-Scale Weak Supervision
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
Port of OpenAI's Whisper model in C/C++
Free and open — grab the source, star the repo, and start creating.