KaniTTS

KaniTTS

Offers the Kani text to speech models from nineninesix in a Gradio app, with several model sizes and English and Chinese variants. It suits developers and creators who want to try a small speech model.

Made by nineninesix 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: KaniTTS is a Gradio app for the Kani text to speech models from nineninesix. The Space names six models, including 370m, 450m and 400m sizes, English and Chinese variants, and an NVIDIA audio codec. The repository holds app.py, two YAML files, helper scripts and a requirements file.

Who it suits: People who want to hear a small text to speech model read their own text, in a browser or on their own machine.

Industry | Who uses it | How they use it | Example
Content | Video creators | Try a voice for narration drafts | A short script read aloud
Software | App developers | Test a small speech model before building on it | A line for a prototype
Language learning | Teachers | Hear text spoken in English or Chinese | A practice sentence
Research | Speech researchers | Compare the named Kani model sizes | The 370m and 400m models

If you need a voice with proven studio quality or documented limits, you can skip it, since the source gives no quality figures.

Where to use it:

• A web browser, on the Space's page at Hugging Face
• Windows, macOS or Linux with Python, from a downloaded copy

Use case | Who | What they get back
Narration drafts | Creators | Spoken audio from typed text
Model comparison | Researchers | Speech from different Kani model versions
Code reading | Developers | A short Gradio app with YAML settings to study

How to set it up:

• Python on your computer
• Python packages from requirements.txt, with Gradio 5.46.0
• At Hugging Face the Space runs on zero-a10g hardware, so a copy may want a similar GPU

1. Install the Hugging Face hub package.
2. Download the Space with the hf command.
3. Move into the downloaded folder.
4. Install Gradio 5.46.0, then the requirements.
5. Start app.py.

pip install -U huggingface_hub
hf download nineninesix/KaniTTS --repo-type space --local-dir KaniTTS
cd KaniTTS
pip install gradio==5.46.0
pip install -r requirements.txt
python app.py

How to use it: Online, open the Space page, type some text and run it. On your own computer, start the app with the command below and use the screen it opens.

python app.py

Tool | What it does
app.py | The main Gradio file that starts the app
examples.yaml | A YAML file that, going by its name, holds sample inputs
model_config.yaml | A YAML file that, going by its name, holds model settings
requirements.txt | The Python packages the app needs

Example: You type a short sentence and run it. The source does not show the screen, but a text to speech app returns audio of the sentence being spoken.

For newcomers and experts: Open the Space page and read one short sentence aloud through it before trying anything long. Hearing a few samples tells you quickly whether the voice suits your use.

The model names carry sizes such as 370m, 400m and 450m, plus an en and a zh variant and pt builds. The Space also names an NVIDIA nemo nano codec at 22 kHz. The source does not explain how the variants differ, so read model_config.yaml and the model pages for that.

You type text in the app
app.py loads a Kani model
The model and codec turn text into audio
The speech plays in the screen

Good to know:

• Hardware: zero-a10g at Hugging Face, built with Gradio 5.46.0
• License: apache-2
• Cost to run: none named
• Data it reaches: the text you enter, plus the named models and codec
• Not for: the source names no limits

On Hugging Face, 13 November 2026

• Likes: 116
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/nineninesix/KaniTTS
• Files: https://huggingface.co/spaces/nineninesix/KaniTTS/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.