Skip to content
Mirayaمرايا
All projects
AIBeta

Qra Lia

Photograph any paper document and get it explained in Moroccan Darija, in text and out loud.

Case study

Draft for review

  1. Problem

    Qra Lia was born at the GOMYCODE × NVIDIA “Come Build with AI” hackathon. The problem it started from: in Morocco, 24.8% of people cannot read, and 51% of those aged 50 and over (HCP, 2024 census). Yet water and electricity bills, bank letters and official mail arrive in French or Modern Standard Arabic, which leads to late fees, successful scams and a constant need for someone else to read every letter.

  2. Solution

    You take a photo of any paper document, and Qra Lia explains it in Moroccan Darija, in text and out loud: what it is, who sent it, how much, by when, what to do, and whether it looks like a scam. The AI reads the document, then fixed code rules check deadlines and scam signs; ID and bank numbers are masked automatically, and no photo is ever stored. The explanation can be played aloud, sent to a family member or turned into a deadline reminder. Built with Next.js, TypeScript and Google Gemini, with Groq as an automatic fallback.

  3. Result

    The project was built in one day during the hackathon (27 September 2026) by a team of three. The team placed 6th in Morocco and 10th in Africa, winning the 6th prize: 8 hours of mentoring with EY Studio+. Next steps: a WhatsApp bot that replies with a Darija voice note, a real Darija voice, and an Android app.