Autocomplete

When you start typing a word, the engine looks at what you've written so far and predicts what comes next. Not just the rest of the current word; it can also suggest the next word in your sentence.

It does this by understanding how Darija words typically follow each other. Think of it like a friend who finishes your sentences because they know the language well enough to guess where you're going.

This tool works differently from the transliterator. The transliterator follows fixed rules to convert between scripts, no guessing involved. Autocomplete, on the other hand, was trained on real Darija text so it can recognize patterns in how people actually write. Two small engines work together behind the scenes: one focuses on the next few letters, the other looks at the bigger picture of your sentence. Together they produce suggestions in milliseconds.

Autocomplete for Darija doesn't really exist anywhere else. Your phone's keyboard doesn't know Darija. Most text prediction tools only understand Modern Standard Arabic or French, which means millions of Moroccans type every day without any help from their devices.

This is a step toward changing that. The goal is for Darija to eventually work everywhere your language should: in keyboards, search bars, messaging apps, and anywhere else people type.