RepoPicks Radar

Updated every day around 23:30 UTC

Guides

VieNeu-TTS

pnnbao97/VieNeu-TTSReviewed by AI

Checked by RepoPicks Radar on

What is VieNeu-TTS?

A Vietnamese text-to-speech program: you paste text and it speaks it with one of 25 built-in voices (northern, central and southern accents), or with a voice copied from a 3-8 second recording. It also does two-speaker conversations and can stream audio to other apps through an OpenAI-style API. Windows has a ready installer; on Mac and Linux you install the uv tool and run a few commands to open a local web page. It runs on a plain processor (about twice real time) and much faster with an NVIDIA card. Text and voices stay on your machine.

VieNeu-TTS is free and open source (Apache-2.0) and can replace ElevenLabs, Murf, and Fish Audio.

How easy is it to install VieNeu-TTS?

B · installer plus one simple step · Windows

  • Windows B
  • macOS C
  • Linux C

On Windows there is a standalone installer in the releases (VieNeu_0.18.22_x64-setup.exe, 90 MB), so you install and open it; the 90 MB file cannot hold the voice models, so the app fetches them from Hugging Face on first use. On Mac and Linux the README says the installer is not ready yet: you install the uv tool, run "uv sync" and "uv run vieneu-web" in a terminal and then open a local web page.

VieNeu-TTS on GitHub

Stars
3,188on Nov 10, 2026
Downloads
1,837127 this week
Growth
+89stars in 6 days

Stars over time

3,1203,1403,1603,1803,200 Nov 6Nov 7Nov 8Nov 9Nov 10 3,1203,1403,1603,1803,200 Nov 6Nov 8Nov 10

Facts

Runs on
Windows, macOS, Linux
You need to install
Nothing extra
To build from source
Python, NVIDIA graphics card
Disk space
90 MB–270 MB · from the installer
Accounts or keys
None
Cost
Free · Everything in this repository is free under Apache-2.0. The newer v4 model is not open source and is offered only through the author's own service, vieneu.io.
License
Apache-2.0 (Permissive)
Last change on GitHub
Nov 9, 2026

VieNeu-TTS risk signals (7)

  • AttentionInstall command runs a script from the internet

    Install command pipes a remote script into a shell: curl -LsSf https://astral.sh/uv/install.sh | sh

    automatic check

  • AttentionDependencies with known vulnerabilities

    Pinned dependencies with known vulnerabilities (OSV): accelerate: GHSA-4j2p-28q2-5m79, PYSEC-2026-3804; aiohttp: GHSA-2fqr-mr3j-6wp8, GHSA-2g87-pp6c-29x5, GHSA-2vrm-gr82-f7m5; anyio: GHSA-5p39-cfhj-2xmp, GHSA-82r6-8w77-94w6, PYSEC-2026-4024; brotli: GHSA-2qfp-q593-8484, GHSA-9v8h-qqgm-4m35, PYSEC-2026-2401; certifi: GHSA-6q48-6c7w-9gw2, PYSEC-2026-4320; datasets: GHSA-379c-qx7v-6h59, PYSEC-2026-3716; filelock: GHSA-69hj-f68j-2879, GHSA-8mx5-7m5m-r77g, GHSA-qmgc-5h2g-mvrw; fsspec: GHSA-27vj-qcqg-25rc, PYSEC-2026-4217; gradio: GHSA-39mp-8hj3-5c49, GHSA-6655-8ph2-63j3, GHSA-7hp7-4p35-3cx2; idna: GHSA-65pc-fj4g-8rjx, PYSEC-2026-215; lxml: GHSA-vfmq-68hx-4jfw, PYSEC-2026-87; msgpack: GHSA-6v7p-g79w-8964, GHSA-j586-36cw-2gc2, PYSEC-2026-3625; multidict: GHSA-54p9-h82j-f925, PYSEC-2026-4244; orjson: GHSA-hx9q-6w63-j58v, PYSEC-2026-107; pillow: GHSA-22hg-r758-rrw9, GHSA-24rc-8hj2-v98c, GHSA-254f-3936-288h; protobuf: GHSA-7gcm-g887-7qv7, PYSEC-2026-1805; pyarrow: GHSA-rgxp-2hwp-jwgg, PYSEC-2026-113; pygments: GHSA-5239-wwwm-4pmq, PYSEC-2026-2987; python-multipart: GHSA-5rvq-cxj2-64vf, GHSA-6jv3-5f52-599m, GHSA-mj87-hwqh-73pj; ray: GHSA-hhrp-gw25-jr43, GHSA-mw35-8rx3-xf9r, GHSA-q5fh-2hc8-f6rq; starlette: GHSA-82w8-qh3p-5jfq, GHSA-86qp-5c8j-p5mr, GHSA-jp82-jpqv-5vv3; transformers: GHSA-29pf-2h5f-8g72, GHSA-69w3-r845-3855, GHSA-fgcw-684q-jj6r

    automatic check

  • Attentionvoice cloning of real people

    README advertises "Instant voice cloning from a 3-8s clip" and the author withheld version 4 from open source "Due to the strength of its voice-cloning capabilities and the potential for misuse"; pyproject.toml shows the audio watermark is optional ("audio watermarking is opt-in via pip install vieneu[watermark]"), so generated audio is normally unmarked.

    AI review

  • Attentionoptional remote code execution setting

    .env.example has VIENEU_TRUST_REMOTE_CODE=1, described as "Chỉ bật nếu tin tưởng repo nhập ở Custom Model: cho phép repo đó chạy code Python riêng" (only turn it on if you trust the repository you type in Custom Model: it lets that repository run its own Python code).

    AI review

  • Infolocal web ui without password by default

    .env.example keeps the login commented out ("# VIENEU_WEB_AUTH=admin:doi-mat-khau-nay"), so the Gradio page on port 7860 and the streaming API have no password unless you set one; this matters if you run it in Docker or on a shared network.

    AI review

  • Infonewest version not open source

    README: "VieNeu-TTS v4 ... is proprietary and will not be open-sourced. It is available exclusively through the VieNeu API / vieneu.io", and "Dubbing, Lecture, Tiktok Reel and Audiobook features are exclusively available in the official VieNeu App".

    AI review

  • Infooptional public tunnel in old docker image

    The deprecated v2 server instructions run "docker run --gpus all -p 23333:23333 ... pnnbao/vieneu-tts:latest --tunnel", which opens a built-in bore tunnel and gives your server a public address on the internet.

    AI review

Evidence is shown in its original language.

What Radar could not check about VieNeu-TTS

  • Whether the Windows installer really works without any extra steps, and how big the voice model download is on first use (the checks count only the 90 MB installer).
  • How close a cloned voice sounds to the original, and how good the English parts of bilingual text are.
  • The speed numbers in the README (first audio in about 115 ms, 16 streams on an RTX 3060) were measured by the author on one machine.
  • Whether the Mac and Linux terminal route works today, since there is no installer for them yet.
  • Whether the app or the installer contacts the author's service (vieneu.io) while you use it.

Similar projects

The RepoPicks app

Coming soon: install in 1 click with the RepoPicks app.

Before you download

See the release page on GitHub