Arabic Dialects · Speech to Text

Iraqi · عراقي

Iraqi Arabic
Speech to Text

An STT model for Iraqi Arabic, understanding the natural Arabic-Persian-Turkish lexical blend across Baghdad, Basra, and Mosul varieties, serving oil & gas, government, and commerce.

dialect: "iraqi"30M+ native speakers · real-time · code-switching

30M+

Native Iraqi Arabic speakers

<11%

WER on Iraqi colloquial speech

<300ms

First-word streaming latency

3+

Iraqi sub-dialects covered

Try it

Speak Iraqi Arabic.
Watch it transcribe.

Record yourself speaking Iraqi Arabic 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 Iraqi Arabic?

Iraqi Arabic is one of the most linguistically distinctive Arabic dialects. It carries the imprint of Mesopotamian civilization in its vocabulary, with a unique blend of Persian and Turkish loanwords that comes from centuries of cultural exchange.

Key features that distinguish Iraqi Arabic:

  • Common Persian vocabulary: "چاي" (chai, tea), "پنجره" (panjara, window), "كاكا" (kaka, brother/friend).
  • Turkish vocabulary: "دولمة" (dolma), "بقلاوة" (baklava), "چلو" (chelo, rice).
  • Baghdadi: the most prestigious and widespread variety in Iraq.
  • Code-switching with English in business settings and the diaspora.

Native speakers

30M+

Iraq

Sub-dialects

3+

Baghdadi · Basrawi · Mosuli

Linguistic blend

Trilingual

Arabic + Persian + Turkish

Hamsa WER

<11%

on colloquial Iraqi audio · 2024

The problem

Why generic Arabic STT fails
on Iraqi Arabic

Generic models reach 38–65% word error rate on colloquial Iraqi speech. Here is why.

Persian and Turkish words break generic models

Models trained only on Standard Arabic fail when they meet the Persian and Turkish words common in everyday Iraqi speech. They treat them as speech errors or drop them entirely, which drives the error rate up sharply.

Wide variation between Baghdadi, Basrawi and Mosuli

The large phonetic differences between Iraqi varieties make a single flat model insufficient. Basrawi carries Gulf features and Mosuli shares traits with Levantine, which confuses generic models not built for Iraq.

Technical terminology in oil and gas

Transcribing technical meetings at energy companies means handling specialized terms spoken with an Iraqi accent. It is a double challenge that a generic model, never trained on both vocabularies together, fails at.

The Iraqi diaspora in multilingual settings

Iraqi speakers in Europe and America frequently drop English words into Iraqi speech, and generic models do not handle this mixing with enough accuracy.

The solution

How Hamsa handles
Iraqi Arabic natively

✓

Training on the full Iraqi lexicon

Training data covering the Arabic-Persian-Turkish vocabulary common in Iraqi speech, with broad representation of informal, business and diaspora speech.

✓

A model covering Baghdadi, Basrawi and Mosuli

Dedicated acoustic models cover the main Iraqi varieties, and each audio segment is routed automatically to the model that best matches its features.

✓

Energy-sector terminology support

A specialized, customizable dictionary for the needs of the oil and gas sector, letting you add technical terms and company names to improve accuracy.

✓

Arabic-English code-switching for the diaspora

A code-switching detection layer recognizes transitions between Arabic and English and processes each segment in its own language, keeping transcripts accurate in diaspora settings.

Capabilities

Everything you need for
Iraqi Arabic STT

✓

Native Iraqi Transcription

Trained on real Iraqi speech across its Baghdadi, Basrawi and Mosuli varieties. Not MSA with a dialect layer on top.

✓

Persian-Turkish-Arabic Vocabulary

The model is trained on the Persian and Turkish loanwords common in everyday Iraqi speech, so it never treats them as errors.

✓

Speaker Diarization

Identify and label individual speakers in meeting and call recordings. Essential for energy companies and government institutions.

✓

Smart Punctuation

Automatic sentence boundaries and punctuation placed correctly for Iraqi grammar and the structure of Baghdadi speech.

✓

Real-time Streaming

Under 300ms first-word latency. Transcript fragments arrive while the speaker is still talking.

✓

Custom Vocabulary

Add your company's energy-sector terms, brand names or government terminology to improve accuracy in your domain.

Transcript examples

Real Iraqi Arabic.
Accurate transcription.

E-commerce contact center

هلا بيك عيوني، الطلبية مالتك راح توصلك باچر الصبح إن شاء الله.

"Welcome, dear. Your order will reach you tomorrow morning, God willing."

Iraqi: "مالتك" (yours), "باچر" (tomorrow, ك pronounced ch), "عيوني" (dear)

Customer service

والله صارلي أسبوع أخابركم وماكو أحد يرد عليّ، شنو القصة؟

"Honestly, I've been calling you for a week and nobody answers. What's going on?"

Natural Iraqi: "أخابر" (to phone), "ماكو" (there is no), "شنو" (what)

Energy sector meeting

هسه الـ shutdown مال المصفى راح يكون الأسبوع الجاي، لازم نخلص الـ inspection قبله.

"The refinery shutdown is now set for next week, so we have to finish the inspection before it."

Code-switching: "shutdown" and "inspection" mid-sentence, "هسه" (now), "مال" (of)

Use cases

Where Iraqi Arabic STT
delivers results

Oil & Gas

Energy company meeting transcription

Converting oil and gas company meetings and reports to accurate text documents, with support for specialized technical terms and Iraq's unique linguistic blend.

Digital Government

eGovernment services

Accurate Iraqi speech recognition for digital government services and citizen interaction platforms, serving people the way they actually speak.

E-Commerce

E-commerce contact center

Customer service call transcription for Iraqi and regional e-commerce platforms, for quality monitoring and analysis.

Financial Services

Banking call documentation

Transcribing banking and financial service calls for compliance and quality purposes, with full support for the Islamic banking terminology used in Iraq.

Accuracy benchmark

High-accuracy STT for Iraqi Arabic

Measured on a held-out test set of colloquial Iraqi speech recordings. Lower word error rate means more words transcribed correctly.

Word Error Rate (WER) · lower is better

Hamsa STT (Iraqi)us11%
Generic Arabic STT38.4%
MSA-trained model48.1%
GPT-4o-transcribe52.6%

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

Architecture

How it works

01

Audio in

Send audio via WebSocket stream or REST upload. Hamsa accepts PCM, MP3, M4A, WAV, and WebM, so it works smoothly with any device or telephony platform.

02

Dialect routing

Hamsa detects Iraqi Arabic features automatically and routes the audio to its dedicated models. When code-switching is detected, Arabic and English segments are processed in parallel.

03

Transcript out

Transcripts stream as JSON events with word-level timestamps, speaker labels, confidence scores, and punctuation. Pipe them straight into your database or application UI.

Developer API

Two lines to Iraqi Arabic speech recognition

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

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

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

stream.on('transcript', (t) => console.log(t.text));
// → "شلونك يا أخي؟" · 00:01.240

Related services

More Hamsa for Iraqi 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 shows most 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 Iraqi and Gulf side by side. No separate integrations, no separate contracts.

FAQ

Iraqi Arabic STT,
frequently asked

Does the model understand Persian and Turkish words in Iraqi Arabic?

Yes. The model's vocabulary covers the Persian and Turkish loanwords common in everyday Iraqi speech, and it does not treat them as errors or unknown segments.

Does the STT model cover Basra and Mosul as well as Baghdad?

Yes. The model is trained on audio data from several Iraqi regions, including Baghdad, Basra and Mosul, and routes each audio segment automatically.

Is it suitable for the oil and gas sector?

Yes. Technical energy-sector terminology is built into the model, and you can customize the dictionary with your company's terms to improve accuracy in specialized contexts.

What word error rate (WER) does Hamsa achieve on Iraqi speech?

Hamsa achieves under 11% WER on colloquial Iraqi speech, compared to 38–65% for generic models. The gap is largest in informal speech and mixed-language contexts.

Can the model be customized with terms specific to my organization?

Yes. You can add a custom dictionary of organization names, technical terms and brand names to improve accuracy in your field.

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.