Darija Project
Building the foundational tools for Moroccan Darija.
We believe in the sovereignty of the Moroccan tech ecosystem. Scaling that ecosystem starts with solving Darija data scarcity, and solving data scarcity starts with understanding our own language.
Why Darija
Darija is spoken by over 30 million Moroccans every day, but it barely exists in the digital world. There are almost no tools, no datasets, no keyboards that truly understand it, no spell checkers. When Moroccans type to each other, their devices don't understand a word of it.
We think that's a problem worth solving. Darija has its own rules, its own way of building words, its own sounds and rhythms. It's not broken Arabic and it's not informal French. It's a full language that millions of people think and communicate in. To truly build technology for it, you need tools that understand those rules. Tools that are extremely fast and highly scalable, so they can reach everyone.
Our approach
We build our tools from deep research into the language itself. How it sounds, how words are formed, how it changes from city to city and generation to generation. We study the patterns Moroccans actually use, not what textbooks say they should use.
The result is technology that is fast, lightweight, and gets better over time. Not through more computing power, but through better understanding of the language and through contributions from native speakers who know it best.
Everything we build works in both directions: Arabizi to Arabic and Arabic to Arabizi. That's how Moroccans actually use their language. Sometimes you need one, sometimes you need the other. The tools should just work, whichever way you need them.