Arabic Dialects · Speech to Text

Libyan · ليبي

Libyan Arabic
Speech to Text

Hamsa STT understands Libyan Arabic as people actually speak it, from Tripolitanian with its Italian vocabulary to Benghazi and the south. Trained on real Libyan speech for oil & gas, telecom, and government.

dialect: "libyan"7M+ speakers · real-time · Tripolitanian, Benghazi & southern

7M+

Libyan Arabic speakers

Oil & Gas

Primary economic driver

<300ms

First-word streaming latency

3

Libyan varieties covered: Tripolitanian · Benghazi · Southern

Coming Soon

Demo launching soon

The interactive demo for this dialect is coming soon. Sign up to be notified when it launches.

Join the waitlist →

Dialect overview

What is Libyan Arabic?

Libyan Arabic is a Maghrebi variety spoken by more than 7 million people, reflecting Libya's position as a bridge between eastern and western Arabic. Tripolitanian speech carries Italian words that entered the everyday lexicon during the Italian colonial period, while southern and interior dialects keep distinct Amazigh and desert features.

Key features that distinguish Libyan Arabic from Modern Standard Arabic:

  • /ق/ → /g/. The qaf is usually a hard G: "قال" (he said) sounds like "gaal".
  • Italian loanwords in everyday Tripoli speech: "ماكينة" (car), "كوجينة" (kitchen), "فرشيطة" (fork).
  • Distinct Amazigh and desert features in southern and interior dialects.
  • Colloquial expressions: "شن حالك؟" (how are you?), "توا" (now), "باهي" (fine/OK), "هلبا" (a lot).

Native speakers

7M+

Libya

Regional varieties

3

Tripolitanian · Benghazi · Southern

Leading sector

Oil & Gas

Primary economic driver

The problem

Why generic Arabic STT fails
on Libyan Arabic

Models trained on MSA struggle with real Libyan speech: Italian loanwords, a hard-G qaf, and strong regional variation. Here is why.

Training data that does not match Libyan speech

Most Arabic STT models are trained on formal MSA content such as news bulletins and read text. Libyans do not speak that way in daily life. Libyan Arabic differs from MSA in pronunciation, vocabulary, and structure, which makes real Libyan speech hard for models that were never trained on it.

Italian vocabulary and technical terms

Tripolitanian speech carries Italian words that have become part of the everyday lexicon, such as "ماكينة" (car) and "كوجينة" (kitchen), while oil and gas conversations are full of English technical terms. Generic models treat these words as errors or drop them, and part of the meaning is lost.

Pronunciation differences

In Libyan Arabic the qaf /ق/ is usually pronounced /g/, so "قال" (he said) sounds close to "gaal", alongside rhythm and pronunciation patterns that differ from MSA. These phonetic differences confuse MSA-trained models, which then write words the speaker never said.

Regional variation

Italian-influenced Tripolitanian differs from the Benghazi variety in the east, and southern and interior dialects keep distinct Amazigh and desert features. A model that ignores these differences treats them as noise, and transcription accuracy drops.

The solution

How Hamsa handles
Libyan Arabic natively

✓

Authentic Libyan training data

Hamsa builds its model on data from Libyan speakers in Tripoli, Benghazi, Sabha, and surrounding regions, not on repurposed MSA data or synthetic speech.

✓

Libyan-specific language model

A language model trained on Libyan vocabulary and structures as they are really used, including everyday expressions such as "شن" (what), "توا" (now), "باهي" (fine/OK), and "هلبا" (a lot), and the Italian words common in Tripoli.

✓

Technical and loanword layer

The model picks out English technical terms and Italian words inside the same Libyan sentence and writes each one in its correct form, without losing the context of the conversation.

✓

Regional variation awareness

One model covers Tripolitanian, Benghazi, and Southern Libyan speech and handles the phonetic differences between them, with no separate setup for each region.

Capabilities

Everything you need for
Libyan Arabic STT

✓

Native Libyan Transcription

Trained on colloquial Libyan speech as it is actually spoken in Tripoli, Benghazi, and the south. Not MSA with a dialect layer on top, but Libyan acoustic and language models from the ground up.

✓

Loanword Recognition

The model recognizes the Italian words rooted in Tripolitanian speech and the English technical terms common in the oil and gas sector, and writes each word in its correct form.

✓

Speaker Diarization

Identify and label each speaker in call and meeting recordings. Essential for contact centers and energy-sector operations teams.

✓

Smart Punctuation

Sentence boundaries, commas, and question marks are added automatically according to the structure of Libyan speech, not MSA grammar rules.

✓

Real-time Streaming

Under 300ms first-word latency. Text arrives while the speaker is still talking, with no waiting for the end of the sentence.

✓

Custom Vocabulary

Add field names, equipment, company names, and technical terms specific to your domain to raise recognition accuracy in your context.

Transcript examples

Real Libyan Arabic.
Accurate transcription.

Telecom call center

باهي يا خوي، توا نشوفلك الفاتورة ونرجعلك بعد شوية.

"OK, brother, I'll check your bill now and get back to you shortly."

Libyan: "باهي" (OK), "توا" (now), "شوية" (a little while)

Customer service

شن صار في الطلب متاعي؟ قعدت نستنى فيه هلبا ومحد رد عليّ.

"What happened to my order? I've been waiting a long time and nobody has answered me."

Libyan: "شن" (what), "متاعي" (my), "هلبا" (a lot)

Oil & gas field call

الـ pump متاع البير واقف من الصبح، لازم نبعتوا technician توا.

"The well pump has been down since the morning. We need to send a technician now."

Technical English: "pump" and "technician" mid-sentence, "متاع" (of)

Use cases

Where Libyan Arabic STT
delivers results

Oil & Gas

Voice Communication in the Energy Sector

Transcribe meetings, field calls, and voice reports in Libyan Arabic for teams that bring international energy companies together with local staff and communities.

Telecom & Contact Centers

Libyan Contact Centers

Transcribe customer calls and monitor service quality in real time, understanding Libyan speech as callers from Tripoli, Benghazi, and the south actually speak it.

GovTech

Government & Citizen Services

Voice interfaces for government services that understand Libyan citizens in their everyday dialect, not formal MSA alone, making digital services easier to reach.

Media & Digital Services

Content & Digital Services

Turn Libyan programs, interviews, and audio content into searchable text, and generate accurate captions for digital platforms and services.

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 Libyan speech features, whether Tripolitanian, Benghazi, or Southern, and routes the audio to the matching acoustic and language models, handling Italian loanwords and English technical terms without losing context.

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 Libyan Arabic speech recognition

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

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

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

stream.on('transcript', (t) => console.log(t.text));
// → "شن حالك يا خوي؟ توا نجيك" · 00:01.840

Related services

More Hamsa for Libyan 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.

✓

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 Libyan and Tunisian. No separate integrations, no separate contracts.

FAQ

Libyan Arabic STT,
frequently asked

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

MSA is the formal register used in media, education, and official documents, while Libyan Arabic is the everyday language of more than 7 million speakers. It differs in pronunciation, such as the qaf pronounced /g/, and in vocabulary, including Italian words in Tripoli and Amazigh features in the south. That is why MSA-trained models clearly struggle with natural Libyan speech.

Does the model support the different Libyan regional varieties?

Yes. The model is trained on data from speakers in Tripoli, Benghazi, Sabha, and surrounding regions, and covers Tripolitanian, Benghazi, and Southern Libyan speech with a single model and no separate setup.

Is the model suited to the oil and gas sector?

Yes. The model handles English technical terms inside Libyan speech and can take custom vocabulary lists covering field names, equipment, and companies, which makes it suited to transcribing meetings, field calls, and voice reports.

Does it work for real-time contact centers?

Yes. Real-time streaming with under 300ms latency is designed for contact centers and voice assistants, with speaker diarization and automatic punctuation.

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.