Kokoclone

Kokoclone

Speaks typed text in a voice cloned from a short reference clip, in eight languages. It runs on a laptop CPU and suits voice assistant prototypes, narration and demos.

Made by Stevensayer7 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: Kokoclone turns typed text into speech in the voice of a short audio clip you give it. It builds on the Kokoro text-to-speech engine and adds voice cloning. It handles eight languages and has a Gradio web page for quick tests.

Who it suits: Developers and creators who need a cloned voice quickly, such as for a voice assistant prototype, a narration or a demo. It also suits researchers running small voice experiments.

Title: Who uses voice cloning
Industry | Who uses it | How they use it | Example
Software | Developers of voice assistants | Prototype a speaking assistant with a chosen voice | Test a reply in a cloned voice
Media | Content creators | Narrate scripts in a voice they own | Read a short script aloud in their own voice
Research | Speech researchers | Run small experiments on voice transfer | Compare a clone against its reference clip
Education | Language teachers | Generate spoken examples in several languages | Make a spoken phrase in French or Japanese

If you need a fixed stock voice with no reference clip, you can skip it.

Where to use it:

• A web browser: the Space runs on Hugging Face with nothing to install
• Your own computer, running Python 3.12.12 on Windows, macOS or Linux
• A standard laptop CPU, or a GPU for faster speech

Use case | Who | What they get back
Clone a voice from a short clip | Creators | Typed text spoken in the voice of the reference clip
Generate speech in another language | Teachers and developers | Speech in English, Hindi, French, Japanese, Chinese, Italian, Portuguese or Spanish
Build a quick demo | Developers | A Gradio page with text, language and reference upload

How to set it up:

• Python 3.12.12 and pip
• The Python packages in requirements.txt
• The system packages in packages.txt, which Hugging Face installs with apt-get

1. Install the Hugging Face hub package and download the Space into a folder.
2. Move into that folder.
3. Install Gradio 6.8.0, then the packages from requirements.txt.
4. Start the app with Python.

pip install -U huggingface_hub
hf download Stevensayer7/clone --repo-type space --local-dir clone
cd clone
pip install gradio==6.8.0
pip install -r requirements.txt
python app.py

The README says NVIDIA users who want GPU speed install the requirements and the GPU build of kokoro-onnx instead.

pip install -r requirements.txt
pip install kokoro-onnx[gpu]

How to use it: Start the app, then open the page it serves in your browser. Type your text, pick a language, upload a reference voice and generate the speech. Model files download on the first run. The README also shows a Python call.

from core.cloner import KokoClone

cloner = KokoClone()

cloner.generate(
    text="This voice is cloned using KokoClone.",
    lang="en",
    reference_audio="reference.wav",
    output_path="output.wav"
)

Example: You upload a reference clip of about three to ten seconds, type a sentence in English and pick the English language option. The app generates that sentence as speech in the voice of your clip.

For newcomers and experts: If you are new to this, open the Space, record a short clip of your own voice and upload it. Type one sentence and listen to the result. Use only voices you have the right to copy.

If you know the field, it pairs the Kokoro ONNX text-to-speech engine with the Kanade tokenizer for voice conversion, which gives zero-shot cloning from one short clip. The trade is a small, fast model that runs on CPU, in exchange for the fidelity a larger system might reach.

You type text and pick a language
You upload a short reference clip
Kokoro speech is converted to your voice
The cloned audio comes back

Good to know:

• Languages: English, Hindi, French, Japanese, Chinese, Italian, Portuguese and Spanish
• First run: the app downloads the model files it needs automatically
• Data it reaches: the reference clip you upload, and model files fetched on first run
• License: apache-2
• Cost to run: none named

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/Stevensayer7/clone
• Files: https://huggingface.co/spaces/Stevensayer7/clone/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.