Arabic Dialects · Text to Speech

MSA · فصحى

Text to Speech
in Modern Standard Arabic

Broadcast-standard MSA voices, formal, educational, and news-quality, for Arab media, e-learning, government services, and official applications across all Arab countries.

dialect: "msa"4 voices · streaming · SSML · diacritics

4.1 / 5

MOS naturalness on MSA speech

4

Broadcast and educational voices

<300ms

First audio chunk latency

MENA

Region-wide official language coverage

Try it

Type Modern Standard Arabic.
Hear it in a broadcast voice.

Enter any MSA text, with or without diacritics, and hear it in a natural voice. The demo uses Hamsa's MSA voice, the same model available through the API.

Text to Speech · text in → speech out

No sign-up
Standard Arabic
78/300

Type anything, pick a voice, and press play.

Free to start · in your browserMake voiceovers in Hamsa Media

Register overview

Synthesizing Modern Standard Arabic

Modern Standard Arabic is the formal, learned register of Arab media, education, courts and government across 22+ countries. It is nobody's mother tongue, but audiences everywhere in the Arab world understand it. For news, e-learning and official services, the voice has to meet broadcast standards: every letter articulated, every vowel where it belongs.

Key features Hamsa TTS reproduces natively:

  • Precise articulation of every letter, including the emphatic consonants and the qaf as /q/.
  • Case endings voiced in formal reading, and word endings dropped at a pause.
  • Numbers and dates read with the gender and case agreement MSA requires.
  • Diacritics that settle ambiguous words: "عِلْم" (knowledge) versus "عَلَم" (flag).

Reach

420M+

Across the MENA region

Hamsa MOS score

4.1/5

naturalness on MSA speech

Streaming latency

<300ms

first audio chunk

Available MSA voices

4

broadcast & educational · female & male

The problem

Why generic Arabic TTS falls short
of broadcast MSA standards

In formal content, listeners catch a misplaced vowel immediately, and trust the voice less.

Missing diacritics confuse pronunciation

Arabic text is usually written without short vowels, so one word can have several readings: "علم" can be "عِلْم" (knowledge), "عَلَم" (flag) or "عَلِمَ" (he knew). Generic engines pick the wrong reading or get the case endings wrong, an error listeners notice immediately in formal contexts.

Monotone delivery unfit for media

News and official programmes call for a balanced delivery with deliberate stress and intonation. Flat, fixed-pitch voices tire listeners in long-form content and cost a bulletin its credibility.

Modern terms and proper names

Arabized scientific and technical terms and the names of foreign people and institutions are mispronounced by engines that are not kept up to date, which hurts the clarity of media and educational content.

Numbers and dates by MSA rules

Reading numbers in MSA follows rules of gender agreement and case. Generic engines read them in a single form or in colloquial Arabic, and the errors show up in financial bulletins and official reports.

The solution

How Hamsa synthesizes
broadcast-standard MSA

✓

Broadcast-standard voices

Hamsa's MSA voices are built to meet the standards of Arab broadcasting: clear articulation, balanced delivery, and stress suited to news, official, and educational content.

✓

Full and partial diacritic support

The engine accepts fully or partially vocalized text, so you control the vowels where precision matters and pass the rest of the text as it is.

✓

Number and terminology processing

A text-processing layer reads numbers and dates according to MSA rules, with a custom lexicon for modern terms and proper names.

✓

Delivery style control

Each voice supports multiple speaking styles, controlled through the `style` API parameter or SSML tags. Use `style: "professional"` for news and official services, or `style: "warm"` for educational content.

Modern Standard Arabic voices

Four broadcast-standard MSA voices

Two broadcast voices and two educational voices. All supporting streaming, SSML, diacritics, and style control.

N
Male

nasser

Broadcast – Formal

Male broadcast-standard voice for news and official programmes.

voice: "nasser-msa"
N
Female

nour

Broadcast – Neutral

Neutral female broadcast voice for news and media content.

voice: "nour-msa"
A
Male

abdulrahman

Educational – Measured

Calm, detailed male voice for e-learning and academic courses.

voice: "abdulrahman-msa"
S
Female

sara

Educational – Clear

Clear, precise female voice for education platforms and training content.

voice: "sara-msa"

Capabilities

Everything you need for
Modern Standard Arabic TTS

✓

Broadcast-standard Pronunciation

Precise articulation and a balanced delivery to the standards of Arab broadcasting, suited to news, official programmes, and educational content.

✓

News and Educational Voices

Four production-ready voices, male and female, in two core styles: broadcast for news and programmes, and educational for courses and academic content.

✓

Full Diacritic Support

Pass fully vocalized text to set every case ending precisely, or add diacritics only to ambiguous words.

✓

Streaming Synthesis

First audio chunk arrives in under 300ms. Stream straight to browser, phone, or telephony systems.

✓

SSML Support

Fine-grained control over pauses, delivery speed and pitch, plus numbers and dates read according to MSA rules.

✓

Custom Lexicon

Set the pronunciation of proper names, institutions, and modern Arabized scientific and technical terms.

Voice examples

Formal Arabic.
Synthesized naturally.

News · Nasser (Male)

أهلًا بكم في نشرة الأخبار، ونبدأ بأبرز عناوين هذا الصباح.

"Welcome to the news bulletin. We begin with this morning's top headlines."

nasser-msa · professional

E-learning · Sara (Female)

في هذا الدرس نتعرّف إلى دورة الماء في الطبيعة ومراحلها الأربع.

"In this lesson we learn about the water cycle in nature and its four stages."

sara-msa · warm

Government service · Nour (Female)

تم استلام طلبكم بنجاح، وسيصلكم إشعار فور اكتمال المعالجة.

"Your request has been received successfully. You will be notified as soon as processing is complete."

nour-msa · professional

Use cases

Where MSA TTS
delivers results

Broadcast Media

News scripts to audio

Producing broadcast-quality news reports and audio clips with professional MSA voices.

E-Learning

Educational audio content

Turning curricula and academic courses into educational audio with clear, formal MSA pronunciation.

Government Services

Official e-service platforms

A clear, formal voice on government e-service platforms to ensure clarity and credibility.

Social Media

Formal audio content production

Creating MSA audio content for social media platforms, institutions, and digital journalism.

Naturalness benchmark

A natural MSA voice at the highest quality

Evaluated by Arab 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)

Hamsa TTS (MSA)us4.1
Generic Arabic TTS2.9
Conventional synthesis2.4
Cloud TTS (Arabic)3.1

Mean Opinion Score rated by Arab listeners from multiple regional backgrounds · 2024 · Higher is better

Architecture

How it works

01

Text in

Send MSA text via REST or WebSocket, fully vocalized, partially vocalized, or plain, with optional SSML markup. A text-processing layer reads numbers, dates and modern terms according to MSA rules before synthesis.

02

MSA synthesis

The selected broadcast or educational voice converts the prepared text to audio with broadcast-standard articulation. Style, speaking rate, and pitch adjust according to the `style` parameter or inline SSML tags.

03

Audio out

MP3, WAV, or Opus audio streams back in real time. First chunk in under 300ms. Pipe directly to WebRTC, telephony, browser audio, or a media player, with no buffering required.

Developer API

Three lines to Modern Standard Arabic speech synthesis

Pass your MSA text, with or without diacritics, pick a broadcast voice, and stream audio. No manual phonetic transcription.

Read the API docs →
msa-tts.tsstreaming
import { Hamsa } from '@hamsa-ai/sdk';

const audio = await hamsa.tts.synthesize({
  text:    'أهلًا بكم في نشرة الأخبار.',
  voice:   'nasser-msa',
  dialect: 'msa',
  style:   'professional',
});

// → streams PCM audio · first chunk <300ms

Related services

More Hamsa for Modern Standard Arabic

Why Hamsa

Built for Arabic from the ground up

✓

Native, not adapted

Hamsa builds Arabic voices from Arabic speech, not an English TTS model fine-tuned for Arabic. Articulation, case endings, and delivery are Arabic from the start.

✓

Enterprise-grade SLA

99.9% uptime, SOC 2 Type II compliance, and MENA data residency. Your text and audio stay in-region.

✓

Fastest Arabic streaming

Under 300ms first-chunk latency in production. Real-time synthesis for interactive services as well as content production.

✓

13+ dialects, one API

Switch from MSA to Saudi, Egyptian or any supported dialect with a single parameter change. No separate integrations, no per-dialect contracts.

FAQ

MSA TTS,
frequently asked

Why choose MSA instead of a local dialect voice?

MSA is the language of media, education, law and official contexts in every Arab country, and Arab audiences everywhere understand it. That makes it the right fit for news bulletins, educational content and government services aimed at more than one country. For consumer apps and everyday conversation, a local dialect voice may feel closer to the user.

Does Hamsa TTS support vocalized (diacritized) text?

Yes. You can pass fully vocalized text to set every case ending precisely, or add diacritics only to ambiguous words. This matters in educational and official content, where every letter must be pronounced exactly as intended.

How fast is Hamsa's Modern Standard Arabic TTS?

Hamsa streams audio in real time. The first audio chunk arrives in under 300ms from the API call, so it works for interactive services as well as content production.

Which MSA voices are available?

Four voices: Nasser and Nour in a broadcast style for news and official programmes, and Abdulrahman and Sara in an educational style for e-learning and training content. Hamsa voices score 4.1 on the Mean Opinion Score (MOS) scale for naturalness.

Does it support SSML and a custom pronunciation lexicon?

Yes. SSML lets you control pauses, delivery speed and pitch, and the pronunciation lexicon lets you set how proper names, institutions, and modern scientific and technical terms are pronounced.

Explore more

13+ Arabic dialects, supported

Partner ecosystem

Trusted Across the
Technology Ecosystem

Hamsa works with leading technology companies, integrators, and solution providers to bring enterprise Voice AI to organizations across the region.