About Sawalni
The language models no one else will build.
Sawalni (سولني — “Ask me” in Moroccan Darija) is the first large language model built natively for Moroccan and North African languages. Not a wrapper of GPT or Claude. A sovereign model with custom tokenization, trained on data we collected, cleaned, generated and classified ourselves.
Our Mission
Close the language gap in AI. 500 million Arabic speakers use dialects invisible to frontier AI. Dozens of African and Amazigh languages lack basic NLP tooling. We build the models that close the gap.
What Sawalni Can Do
- →Chat in Darija, Amazigh, Arabic, French, English, or Spanish — and get answers that actually understand your context.
- →Search the web with AI-powered summaries and source citations.
- →Check weather, currency conversion, world clocks, and more via built-in tools.
- →Read the world's news translated into Darija — your language, your terms.
Languages & Scripts
Sawalni natively supports 3 scripts and 11 language varieties.
عربي
Darija (Arabic)
كيداير؟
Latin
Darija (Arabizi)
Kidayr?
عربي
Hassaniya
كيف حالك؟
Latin
Tachelhit (Latin)
Manzakin?
ⵜⵉⴼⵉⵏⴰⵖ
Tachelhit (Tifinagh)
ⵎⴰⵏⵣⴰⴽⵉⵏ?
Latin
Tarifit (Latin)
Mamec tettiliḍ?
ⵜⵉⴼⵉⵏⴰⵖ
Central Atlas Tamazight
ⵎⴰⵎⴽ ⵜⵜⵉⵍⵉⴷ?
عربي
Modern Standard Arabic
كيف حالك؟
Latin
French
Comment ça va?
Latin
English
How are you?
Latin
Spanish
¿Cómo estás?
Sovereign AI
Sawalni is not a GPT wrapper. Every component — from tokenization to training data to inference — was built from scratch for the languages and cultures of North Africa.
Our custom tokenizer handles Arabic, Latin, and Tifinagh scripts in a single vocabulary. Our training data was collected, cleaned, and classified using our own language identification model (Gherbal). The result is an AI that doesn't just translate — it understands.
The Omneity Ecosystem
Founded by Omar Kamali
Berlin + Salé. Omar grew up speaking Darija, studied in French, works in English and German — and watched as AI after AI launched without understanding a single Moroccan sentence. So he built one that does.
Read the Full Story
From Wikipedia obsession to the first Moroccan LLM — the entire journey, with videos, demos, and milestones.
Our Story →