Arabic Dialects · Speech to Text

Egyptian · Masri

Egyptian Arabic
Speech to Text

Hamsa STT understands real Egyptian Arabic, not normalized Modern Standard Arabic. Purpose-trained on native Masri speech for call centers, media monitoring, and voice apps.

dialect: "egyptian"100M+ native speakers · real-time · code-switching

100M+

Native Egyptian Arabic speakers

<18%

WER on colloquial Masri

<300ms

First-word streaming latency

4

Regional sub-dialects covered

Try it

Speak Egyptian Arabic.
Watch it transcribe.

Record yourself speaking Masri and watch the words appear as you talk. The demo runs Hamsa's full production STT, the same model your API calls will use.

Speech to Text · speech in → text out

No sign-up
Microphone
press to record

Your words will appear here.

Press record and speak Arabic or English. The words appear as you talk.

Free to start · in your browserTranscribe in Hamsa Media

Dialect overview

What is Egyptian Arabic (Masri)?

Egyptian Arabic, called Masri (مصري) or Aamiyya Masriyya, is the spoken vernacular of over 100 million people in Egypt. Decades of Egyptian cinema, music, and television have made Masri intelligible to Arabic speakers across the entire region, making it the de facto pan-Arab dialect of popular culture.

Key features that distinguish Masri from Modern Standard Arabic:

  • /ق/ → /ʔ/، The qaf becomes a glottal stop. "قلت" (I said) sounds like "ʔolt".
  • /ج/ → /g/، The jim is a hard G. "جاي" (coming) sounds like "gaay".
  • Mediterranean loanwords: "أوضة" (room, from Coptic), "طرابيزة" (table, from Italian).
  • Colloquial expressions: "عامل إيه؟" (how are you?), "خلاص" (done/OK), "يعني" (basically).

Native speakers

100M+

Egypt

Regional sub-dialects

4

Cairo · Alexandria · Sa'idi · Delta

Code-switch frequency

~30%

of urban professional speech

Hamsa WER

<18%

on colloquial Egyptian audio · 2024

The problem

Why generic Arabic STT fails
on Egyptian Arabic

A model trained on MSA achieves 40–65% word error rate on Egyptian call center audio. Here is why.

MSA training data mismatch

Most Arabic STT models train on broadcast news, Quran recitation, and formal written text in Modern Standard Arabic, a register no Egyptian speaks conversationally. Real Masri diverges completely in pronunciation, grammar, and vocabulary.

Code-switching blind spots

Educated Egyptians routinely switch between Arabic, English, and French within the same sentence. Generic models treat these transitions as errors or silently drop the non-Arabic words.

Pronunciation divergence

Egyptian Arabic converts /ق/ to a glottal stop and /ج/ to a hard G. A model trained on MSA hears "قلت" and transcribes a word nobody says, breaking meaning at the phoneme level.

Speed and sub-dialect variation

Cairene Arabic is among the fastest Arabic varieties, with reduced vowels and frequent elision. Cairo, Alexandria, Upper Egypt, and Delta speech each have distinct features that flat models collapse into noise.

The solution

How Hamsa handles
Egyptian Arabic natively

✓

Native Egyptian corpus

Millions of hours of real Masri conversational speech, call center recordings, social media audio, Egyptian TV, not resampled MSA or text-to-speech data.

✓

Dialect-specific language model

The language model is trained on Egyptian colloquial vocabulary, grammar patterns, and internet speech, including slang and informal expressions absent from any MSA dictionary.

✓

Code-switching layer

A parallel detection pass identifies language switches and applies the correct decoder for Arabic and English/French segments, transcribing each in its native form.

✓

Regional sub-dialect awareness

Separate acoustic models for Cairene, Alexandrian, Sa'idi (Upper Egyptian), and Delta dialects. The system routes each utterance to the best-matched model automatically.

Capabilities

Everything you need for
Egyptian Arabic STT

✓

Native Masri Transcription

Purpose-trained on Egyptian colloquial speech. Not MSA with a dialect layer on top, native Masri acoustic and language models, from the ground up.

✓

Code-Switch Detection

Egyptian speakers mix Arabic with English and French mid-sentence. Hamsa handles these transitions without losing context or misidentifying the language.

✓

Speaker Diarization

Identify and label individual speakers in call recordings. Critical for Egyptian contact centers processing thousands of hours of audio per day.

✓

Smart Punctuation

Automatic sentence boundaries, commas, and question marks inserted correctly for Egyptian Arabic syntax, not MSA grammar rules.

✓

Real-time Streaming

Under 300ms first-word latency. Transcript fragments arrive as the speaker talks. No silence-detection delay, no waiting for utterance boundaries.

✓

Custom Vocabulary

Add Egyptian-specific brand names, medical terminology, or industry jargon to improve recognition accuracy for your domain.

Transcript examples

Real Egyptian Arabic.
Accurate transcription.

Call center (BPO)

أيوه حضرتك، طيب أنا هعمل ليك ريست للباسورد دلوقتي.

"Yes, I'll reset your password for you right now."

Masri: "أيوه" (yes), "دلوقتي" (now), "ريست" (reset، English loan)

Customer service

والله الموضوع ده بقى مزعلني جداً، أنا كلمتكم المرة الجاية وقالولي خلاص اتحل.

"Honestly, this has been upsetting me. I called last time and they said it was resolved."

Natural Masri: "والله" (honestly), "الجاية" (the next/last), "خلاص" (done/resolved)

Media interview

إحنا بقينا نتكلم عن sustainability بشكل مختلف خالص، يعني الناس بقت واعية أكتر.

"We've started talking about sustainability differently now, people are much more aware."

Code-switching: "sustainability" inserted mid-sentence, "يعني" (basically/like)

Use cases

Where Egyptian Arabic STT
delivers results

BPO & Call Centers

Egyptian Contact Centers

Egypt hosts the largest BPO industry in the Arab world, serving Saudi Arabia, the UAE, and Europe. Hamsa enables real-time quality assurance, transcription, and agent coaching at scale.

Media & Broadcasting

Egyptian Media Monitoring

Egyptian TV and streaming content reaches the entire Arab world. Hamsa auto-generates subtitles, searchable transcripts, and topic classifications for broadcast and digital media.

EdTech

Education & E-learning

Transcribe Egyptian Arabic lectures for searchable archives, generate accessible captions for students with hearing impairments, and enable voice-driven interactive tutoring.

GovTech

Government & Citizen Services

Egypt's Vision 2030 digitization programs require voice interfaces that understand how citizens actually speak, not textbook Arabic spoken by no one outside a classroom.

Accuracy benchmark

The most accurate Egyptian Arabic STT available

Measured on a held-out benchmark of Egyptian call center and conversational audio. Lower word error rate means more words transcribed correctly.

Word Error Rate (WER) · lower is better

Hamsa STT (Masri)us18%
Generic Arabic STT38.4%
MSA-trained model48.1%
GPT-4o-transcribe52.6%

Evaluated on Egyptian colloquial speech benchmarks · 2024 · Lower is better

Architecture

How it works

01

Audio in

Send audio via WebSocket stream or REST upload. Accepts PCM, MP3, M4A, WAV, and WebM, whatever your microphone, telephony platform, or recording system produces.

02

Dialect routing

Hamsa detects Egyptian Arabic features and routes to the Masri-specific acoustic and language models. If code-switching is detected, the parallel Arabic/English/French decoder activates.

03

Transcript out

Transcripts stream as JSON events with word-level timestamps, speaker labels, confidence scores, and punctuation. Pipe directly into your database, UI, or downstream NLP pipeline.

Developer API

Two lines to Egyptian Arabic speech recognition

Drop in the Hamsa SDK, set dialect: "egyptian", and start streaming transcripts. No preprocessing pipeline, no normalization to MSA, no fallback surprises.

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

const stream = hamsa.stt.stream({
  model:       'hamsa-stt-s4',
  dialect:     'egyptian',
  diarization: true,
});

stream.on('transcript', (t) => console.log(t.text));
// → "عامل إيه يا صاحبي؟" · 00:01.840

Related services

More Hamsa for Egyptian Arabic

Why Hamsa

Built for Arabic from the ground up

✓

Native, not adapted

Hamsa trains on real Arabic dialect data, not English models adapted for Arabic. The difference is 25–50 WER points on colloquial speech.

✓

Enterprise-grade SLA

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

✓

Fastest Arabic streaming

Under 300ms first-word latency in production. Real-time transcription that keeps pace with human speech rates.

✓

13+ dialects, one API

Switch dialect with a single parameter. Run A/B tests across Egyptian and Gulf. No separate integrations, no separate contracts.

FAQ

Egyptian Arabic STT ·
frequently asked

What is the difference between Egyptian Arabic and Modern Standard Arabic?

Modern Standard Arabic (MSA) is a formal written register used in pan-Arab media, education, and official documents. It is nobody's mother tongue. Egyptian Arabic (Masri) is the spoken vernacular of 100 million Egyptians, with a distinct vocabulary, grammar, and pronunciation system. STT models trained on MSA fail on real Egyptian speech because the two varieties differ as much as written Latin from spoken Italian.

Does Hamsa handle Egyptian Arabic code-switching with English or French?

Yes. Code-switching, mixing Arabic with English or French within a sentence, is a natural feature of educated Egyptian speech, especially in business and technology contexts. Hamsa detects language switches in real time and applies the correct decoder for each segment, transcribing mixed utterances accurately rather than forcing everything into one language.

What word error rate does Hamsa achieve on Egyptian Arabic?

Hamsa achieves under 18% WER on Egyptian colloquial Arabic benchmarks, compared to 38–65% for generic Arabic STT models. The gap is largest on informal speech, call center audio, social content, unscripted conversation, exactly the data most enterprises need to process.

Can Hamsa handle different Egyptian regional accents?

Yes. Egyptian Arabic has significant regional variation: Cairene (the prestige variety), Alexandrian, Upper Egyptian (Sa'idi), and Delta dialects each have distinct phonological features. Hamsa's acoustic models cover all major Egyptian sub-dialects and route each utterance to the best-matched model automatically.

What industries use Egyptian Arabic STT most?

The largest use cases are BPO and contact centers (quality assurance, real-time transcription, agent coaching), media and broadcasting (subtitling, content indexing, topic detection), fintech (voice verification, call recording compliance), and e-commerce (customer feedback analysis, voice search). Egypt's large informal economy also creates demand for voice interfaces that reach users with limited literacy.

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.