Arabic Dialects · MSA
MSA · فصحىModern Standard Arabic.
The language of Arabic media, education, law, and courts in over 22 countries. MSA differs fundamentally from all spoken dialects — it is a learned and consciously produced language, making building an accurate STT and TTS model for it distinct and precise in its requirements. Hamsa builds a model specifically for this purpose.
dialect: "msa"420M+
Modern Standard Arabic users
MENA
Region-wide official language coverage
<14%
WER on Modern Standard Arabic recognition
Media & Law
Primary official use context
Try it
Meet the Modern Standard Arabic voices
Four of Hamsa’s Modern Standard Arabic voices. Press one to hear it.
Voices
No sign-upEach voice plays a short sample.
Overview
What is Modern Standard Arabic?
Modern Standard Arabic is the unified, standardized form of Arabic used in media, education, legal, and formal contexts in all Arab countries. No one speaks it as a mother tongue in daily home life — it is a learned language — but it is consciously produced by broadcasters, judges, teachers, politicians, and activists across the Arab world.
This learned character makes MSA fundamentally different in its speech challenges: while a dialect model deals with spontaneity and natural variation of everyday speech, an MSA model deals with the diversity of speaker accents from multiple dialects — each pronouncing MSA with their own regional background influencing sounds and stress.
MSA is used extensively in broadcast media, e-learning, courts, official documents, and politician and activist speeches on social media. Hamsa provides a model trained on the wide spectrum of spoken MSA across multiple dialect backgrounds.
Usage scope
420M+
Across the MENA region
Primary context
4
Media · Education · Law · Official
Unique feature
Learned language
Affects pronunciation across multiple backgrounds
WER
<14%
Hamsa benchmark 2024
Hamsa for Modern Standard Arabic
Choose your product
MSA Speech to Text
Modern Standard Arabic speech-to-text, handling the diversity of speaker accents from different dialect backgrounds pronouncing MSA with their own regional influence.
- Under 14% WER on real MSA speech
- Wide range of regional accents covered
- Sub-300ms real-time streaming
- Suited for media, education, law, and official use
MSA Text to Speech
Natural MSA TTS with broadcast and formal voices at high media standards for media, e-learning, and official services.
- Broadcast-standard MSA pronunciation
- Multiple broadcast, formal, and educational voices
- Suited for media, education, and government services
- SSML support with full tashkeel



