ZONOS3 Mini

ZONOS3 Mini

Turns up to 600 characters of text into spoken audio with Zyphra's ZONOS3-Mini model, in several languages and preset voices. A Gradio app that runs in the browser on Hugging Face.

Made by fffiloni on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This is a small Gradio app for Zyphra's ZONOS3-Mini text-to-speech model. You type up to 600 characters, pick a language and a preset voice, and press one button to get a spoken WAV file. The interface was designed by Sylvain Filoni, and it runs as a Hugging Face Space on ZeroGPU.

Who it suits: People who want a short spoken clip from text without setting up a model, and developers who want a worked example of a polished Gradio text-to-speech front end.

Industry | Who uses it | How they use it | Example
Media | Video creators | Turn a short script line into a voice clip | A spoken intro line for a clip
Education | Language learners | Hear a sentence read in another language | A French phrase read aloud
Software | App developers | Study how the audio player and settings are built | Reuse the waveform display idea
Research | Speech model fans | Try the model with different sampling settings | Compare two seeds on one sentence

Skip it if you need long audio, since the text box stops at 600 characters.

Where to use it:

• A web browser, on the Space's page at Hugging Face
• Your own computer, as a local copy, once Python 3.12 is installed
• A Hugging Face ZeroGPU setup, which is where the hosted version runs

Use case | Who | What they get back
Read a short text aloud | Creators and learners | A WAV file that plays in the browser and can be downloaded
Try other languages | Language learners | Speech in the language you chose
Clone-style voice from a clip | Advanced users | Speech guided by an optional reference clip

How to set it up:

• Python 3.12 and pip, for a local copy
• The Python packages listed in requirements.txt
• The system packages listed in packages.txt, which Hugging Face installs with apt-get on its own servers
• Nothing at all if you only use the online Space

1. Open the Space's page at Hugging Face to use it online.
2. For a local copy, install the Hugging Face hub package and download the Space.
3. Move into the new folder and install Gradio at the pinned version.
4. Install the packages from requirements.txt.
5. Start the app with Python.

pip install -U huggingface_hub
hf download fffiloni/ZONOS3-Mini --repo-type space --local-dir ZONOS3-Mini
cd ZONOS3-Mini
pip install gradio==6.30.0
pip install -r requirements.txt
python app.py

The hosted Space runs on zero-a10g hardware. The README does not say what a local run needs beyond the packages above.

How to use it: Write your text, choose a language and a preset voice, then select Generate speech. The finished audio plays on its own when Auto-play result is on, which is the default, and you can turn it off in Voice settings. You can also seek and download the file.

Tool | What it does
Generate speech | Makes the audio from your text, language and voice
Auto-play result | Plays the finished audio without a click
Voice options | Holds the reference clip, consent checkbox, seed, speaking rate and sampling
Examples | Nine buttons that fill in text, language and voice

Example: You pick English and a preset voice, type a short welcome message and select Generate speech. The app shows its progress, then plays a WAV of your message with a waveform line that follows the sound. You can download the file.

For newcomers and experts: If you are new, open the Space online and try one of the nine example buttons. They fill the text, language and voice for you, so the first run takes one click on Generate speech.

If you know the field, each sampling control pairs a slider with an exact number box. They cover temperature, repetition penalty, top-p and top-k, and you also get seed, normalization and speaking rate. The README says the checks were short French and English runs and that it is not an audio-quality benchmark. A reference clip is optional and comes with a consent checkbox, so use only voices you have the right to use.

Type text and choose language and voice
Select Generate speech
ZONOS3-Mini runs on ZeroGPU
The WAV plays in the browser
You download the file

Good to know:

• License: mit
• Cost to run: none named, though the hosted Space uses ZeroGPU quota at Hugging Face
• Data it reaches: the text you type and any reference clip you add, plus model files pulled from Hugging Face
• Not for: long texts above 600 characters, and voices you have no right to copy

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: mit

Links

• Space: https://huggingface.co/spaces/fffiloni/ZONOS3-Mini
• Files: https://huggingface.co/spaces/fffiloni/ZONOS3-Mini/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.