MOSS TTS Nano

MOSS TTS Nano

A voice cloning demo that reads your text aloud in a voice taken from a reference audio clip or a built-in preset. Built by OpenMOSS-Team and run on Hugging Face ZeroGPU.

Made by OpenMOSS-Team 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: MOSS TTS Nano is a text-to-speech demo that does one job, voice cloning. You give it a reference voice, either an audio file you upload or a built-in preset, and it speaks your text in that voice. It is made to run on Hugging Face ZeroGPU.

Who it suits: It suits people who want to hear how a compact voice cloning model sounds before they build with it.

Title: Who might use this demo
Industry | Who uses it | How they use it | Example
Speech research | Researchers | Listen to voice cloning from a small model | Clone one reference clip and judge the result
Software | Developers | Test cloned speech for an app idea | Try a built-in preset voice with a sample text
Audio production | Creators | Hear a script read in a chosen voice | Paste a few lines and generate the audio

If you need plain text-to-speech without cloning, or controls for switching between CPU and other backends, you can skip it.

Where to use it: You can use it online or run a copy yourself.

• Web browser: the Space's page on Hugging Face, with nothing to install
• Windows, macOS or Linux: a local copy that runs on Python 3.10
• Hugging Face Spaces: where the hosted demo runs, on ZeroGPU hardware

Title: What the demo is good for
Use case | Who | What they get back
Clone a voice from a clip | Developers | The full output audio, generated once decoding finishes
Try a preset voice | Anyone curious | Spoken audio in one of the built-in voices
Replay an example | Testers | Text and prompt audio filled in from a built-in example case

How to set it up: The hosted demo needs nothing. A local copy needs Python 3.10 and pip.

• Python 3.10, with pip
• The Python packages that requirements.txt lists

1. Install the huggingface_hub package.
2. Download the Space's files into a folder.
3. Move into that folder.
4. Install Gradio 6.5.1.
5. Install the packages from requirements.txt.
6. Start the app with Python.

pip install -U huggingface_hub
hf download OpenMOSS-Team/MOSS-TTS-Nano --repo-type space --local-dir MOSS-TTS-Nano
cd MOSS-TTS-Nano
pip install gradio==6.5.1
pip install -r requirements.txt
python app.py

How to use it: Open the Space online, or start your copy with the command below. Then pick a reference voice by uploading a clip or choosing a preset, enter your text and generate.

python app.py

Title: The parts you will meet
Tool | What it does
Reference audio upload | Lets you supply the voice to clone
Preset voices | Built-in voices you can pick instead of uploading
Example cases | Fill in the text and show the prompt audio that will be used
app.py | Starts the Gradio page when you run it with Python

Example: You upload a short clip of a voice, type a sentence you want spoken, and press generate. When decoding finishes, the full audio of your sentence comes back in that voice.

For newcomers and experts: If you are new to this, start with a built-in preset voice and a short sentence, so you need no audio file. Use a reference clip only from a voice you have the right to use.

If you know the field, the demo supports only voice cloning and keeps the device on auto, so the page has no switches for CPU use or for the backbone. It names the MOSS-Audio-Tokenizer-Nano model, and the repository ships text normalization code and a weights folder.

Pick a preset or upload a clip
Type the text to speak
The nano model clones the voice
The full audio arrives after decoding

Good to know: The demo is narrow on purpose, and the page hides engine choices.

• Mode: voice cloning only, with no plain text-to-speech mode
• Device: it stays on auto, and the page has no CPU switch
• Output: you receive the full audio after decoding finishes
• License: apache-2
• Cost to run: none named
• Data it reaches: the reference audio you upload and the text you type, plus the Python packages pip installs from requirements.txt
• Not for: choosing CPU execution or a different backbone from the page

On Hugging Face, 13 November 2026

• Likes: 79
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/OpenMOSS-Team/MOSS-TTS-Nano
• Files: https://huggingface.co/spaces/OpenMOSS-Team/MOSS-TTS-Nano/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.