TTS Tounsi

TTS Tounsi

Turns text written in Tunisian Arabic dialect into speech with a fine-tuned XTTS v2 model, using a built-in male or female voice or one you upload. For anyone who needs spoken Tunisian.

Made by Nouira-Oussema 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 Space turns text written in Tunisian Arabic dialect into spoken audio. It uses XTTS v2 from Coqui TTS, fine-tuned on Tunisian dialect recordings. You choose a reference voice, and the speech is made to follow it.

Who it suits: It suits people who need to hear Tunisian dialect read aloud, such as creators who voice scripts and learners who want to hear written text. Developers can also read how a Gradio page wraps a speech model.

Industry | Who uses it | How they use it | Example
Media | Video and audio creators | Voice a short script in Tunisian Arabic | Turn a line of narration into a clip you can listen to
Education | Teachers and learners | Hear how written Tunisian dialect sounds | Check the sound of a sentence before using it in class
Research | Speech researchers | Try a dialect fine-tune of XTTS v2 | Compare the male, female and uploaded reference voices
Software | Developers | Read app.py and the sample voice files | See how a speech model is wrapped in a Gradio page

If your text is in another language or dialect, skip it, since the model was tuned for Tunisian.

Where to use it: The page runs in two places.

• A web browser: open the Space page on Hugging Face, with nothing to install.
• Python 3.11 on Windows, macOS or Linux: start your own copy of the app there.

Use case | Who | What they get back
Voice a script | Creators | Speech in Tunisian dialect that follows the voice you picked
Hear a sentence | Learners and teachers | A spoken version of text written in Tunisian Arabic
Use your own voice as the guide | Anyone with a short recording | Speech shaped by an audio file of 6 to 30 seconds that you upload

How to set it up: The online Space needs none of this. For a local copy you need a few things first.

• Python 3.11
• pip, to install packages
• A terminal on Windows, macOS or Linux

These are Hugging Face's steps for a local copy.

1. Install the Hugging Face hub package with pip.
2. Download the Space's files into a folder named TTS-Tounsi.
3. Move into that folder.
4. Install Gradio 5.0.0, then the packages listed in requirements.txt.
5. Start the app.

pip install -U huggingface_hub
hf download Nouira-Oussema/TTS-Tounsi --repo-type space --local-dir TTS-Tounsi
cd TTS-Tounsi
pip install gradio==5.0.0
pip install -r requirements.txt
python app.py

How to use it: Open the Space page, or start your local copy with the command below. Type your text in Tunisian Arabic, pick a reference voice, then press the button that generates the voice and listen to the result.

python app.py

Tool | What it does
Text box | Takes your text in Tunisian Arabic
Reference voice | Lets you choose the male voice, the female voice or your own uploaded audio
Générer la voix | Makes the speech so you can play it

Example: You type a short sentence in Tunisian Arabic and choose the female voice. You press the generate button, and the page plays back your sentence as speech in that voice. To use your own voice instead, you upload a recording of 6 to 30 seconds.

For newcomers and experts: Text to speech means a computer reads written words aloud. Start by choosing one of the two ready-made voices, typing a short sentence in Tunisian Arabic and playing the result. Only upload a voice you have the right to use.

Underneath, the Space uses a fine-tune of XTTS v2 from Coqui TTS, trained on Tunisian dialect audio. The reference audio steers the voice, with a 6 to 30 second clip, and the output is sampled at 24 kHz. The repository ships sample files for the male and female voices, and the Space runs on cpu-basic hardware at Hugging Face. The source does not say how long a run takes.

You type text in Tunisian Arabic
You pick a male, female or uploaded voice
The fine-tuned XTTS v2 model makes the speech
You play the audio the page returns

Good to know: A few facts are worth checking first.

• Python: the Space and the local steps use Python 3.11.
• Voice cloning: the page can follow a voice you upload, so use only voices you have the right to use.
• License: mit, as the Space's listing shows. The base model comes from Coqui TTS, and the source does not state its license.
• Cost to run: none named. The source lists no API key, account or paid service.
• Data it reaches: your text and any audio file you upload, plus Hugging Face for the Space's files and the models. The source does not say where processing happens.
• Not for: other languages or dialects. The source states no further limits or forbidden uses.

On Hugging Face, 13 November 2026

• Likes: 5
• Built with: gradio
• License: mit

Links

• Space: https://huggingface.co/spaces/Nouira-Oussema/TTS-Tounsi
• Files: https://huggingface.co/spaces/Nouira-Oussema/TTS-Tounsi/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.