AI Model · Text to speech
omni · v3Hamsa-TTS-Omni v2.8.
Expressive, human-like voices in 13+ Arabic dialects, plus English. Fast enough to speak live.
model="hamsa-tts-omni-v2.8"24kHz · streaming ✓ · cloning ✓13+
Arabic dialects, native delivery
<150ms
to the first byte of audio
2
languages: AR · EN
24kHz
streaming studio quality
Try it
Type Arabic.
Hear it back.
Pick a dialect, write anything, and press play. The sheet reads along as it speaks.
Text to Speech · text in → speech out
No sign-upType anything, pick a dialect and a voice, then press play.
Coverage
Eighteen dialects.
Every one native.
Plus English in all accents, for the region's real households, call centers, and audiences. Every card opens that dialect's own page.
Why it sounds human
Not read aloud.
Performed.
For developers
One call.
Any voice.
Stream audio as it's synthesized, or render whole scripts in batch.
const audio = await hamsa.tts.speak({
model: 'hamsa-tts-omni-v2.8',
voice: 'layla', dialect: 'khaleeji',
text: 'مرحباً! كيف أقدر أساعدك اليوم؟',
});
audio.pipe(speaker); // first byte in <150msIn their words
What people say
about Hamsa.
Hamsa's transcription turned our podcasts into high-quality web articles, saving us time and cutting production costs.
Abed Abu Shihade


















