Arabic Dialects · Text to Speech
Tunisian · تونسيTunisian Arabic
Text to Speech
Hamsa TTS speaks Tunisian the way people actually speak it, with all its speed, rhythm, and distinctive vowel deletion. Authentic Tunisian voices, mixed-text support, and under 300ms streaming.
dialect: "tunisian"4 voices · streaming · SSML4.1 / 5
MOS naturalness on Tunisian speech
4
Native Tunisian Arabic voices
<300ms
First audio chunk latency
12M+
Native Tunisian Arabic speakers
Demo launching soon
The interactive demo for this dialect is coming soon. Sign up to be notified when it launches.
Join the waitlist →Dialect overview
Synthesizing Tunisian Arabic
Tunisian Arabic is one of the fastest and most phonologically dense Arabic dialects, spoken by more than 12 million people. Layered on top is a multilingual heritage of French, Italian, Amazigh, and Turkish. For TTS serving Tunisian users, in tourism, telecom, or digital services, the voice has to sound like Tunisia, not a slowed-down newsroom reading.
Key phonological features Hamsa TTS reproduces natively:
- Fast speech rate with systematic short-vowel deletion and syllable compression.
- Tunisian stress and intonation patterns, not MSA broadcast cadence.
- French and Italian words pronounced naturally inside Tunisian sentences.
- Colloquial expressions: "برشا" (a lot), "توا" (now), "باهي" (OK) spoken naturally.
Native speakers
12M+
Tunisia
Hamsa MOS score
4.1/5
naturalness on Tunisian speech
Streaming latency
<300ms
first audio chunk
Available Tunisian voices
4
female & male · multiple styles
The problem
Why generic Arabic TTS sounds
unnatural to Tunisian users
Tunisian listeners notice a slow, formal MSA voice immediately, and trust it less.
The solution
How Hamsa synthesizes
authentic Tunisian speech
Tunisian Arabic voices
Four native Tunisian voices
All trained on native Tunisian speakers. All supporting streaming, SSML, and style control.
Capabilities
Everything you need for
Tunisian Arabic TTS
Voice examples
Real Tunisian Arabic.
Synthesized naturally.
Use cases
Where Tunisian Arabic TTS
delivers results
Naturalness benchmark
Natural Tunisian Arabic speech at the highest quality
Evaluated by native Tunisian listeners on a 5-point Mean Opinion Score (MOS) scale. Higher score means more natural-sounding speech.
MOS Naturalness Score · higher is better (max 5.0)
Mean Opinion Score rated by native Tunisian listeners · 2024 · Higher is better
Architecture
How it works
Developer API
Three lines to Tunisian Arabic speech synthesis
Pass your Tunisian Arabic text, pick a Tunisian voice, and stream audio. No phonetic transcription, no MSA normalization step, no dialect adapter config.
Read the API docs →import { Hamsa } from '@hamsa-ai/sdk';
const audio = await hamsa.tts.synthesize({
text: 'أهلاً بيك، شنية الأخبار؟',
voice: 'mounir-tn',
dialect: 'tunisian',
style: 'warm',
});
// → streams PCM audio · first chunk <300msRelated services
More Hamsa for Tunisian Arabic
Why Hamsa
Built for Arabic from the ground up
FAQ
Tunisian Arabic TTS:
frequently asked
Does Hamsa TTS sound like a real Tunisian speaker?
Yes. Hamsa's Tunisian Arabic voices are trained on native Tunisian speakers and use a voice model designed specifically for Tunisian. The output has authentic Tunisian speed and rhythm and natural colloquial pronunciation, including the vowel deletion and distinctive stress that define how Tunisians actually speak.
What is the difference between Tunisian TTS and generic Arabic TTS?
Generic Arabic TTS engines are trained on Modern Standard Arabic, so when they process Tunisian text they may apply MSA pronunciation, and the voice sounds slow, formal, and unnatural. Hamsa trains on authentic Tunisian speech data with a dedicated voice model, producing speech that reflects real Tunisian delivery at its full speed and rhythm.
How fast is Hamsa's Tunisian Arabic TTS?
Hamsa streams audio in real time. The first audio chunk arrives in under 200ms from the API call, so IVR systems, voice assistants, and tourism apps can start speaking almost instantly with no noticeable delay.
Can I control the speaking style or emotion?
Yes. Each voice supports multiple speaking styles via the `style` API parameter or SSML tags. Use `style: "professional"` for IVR, `style: "warm"` for customer service, or `style: "expressive"` for consumer and tourism apps.
Does Hamsa TTS handle mixed Tunisian and French text?
Yes. A dedicated layer processes Tunisian text before synthesis, including French and Italian words written in Arabic script, numbers, dates, and informal spellings. You can pass Tunisian text as it is, directly.
Explore more



