August 20, 2026, (Inside AI) — A blind Egyptian entrepreneur has built an AI-assisted application that narrates the visual world for people with vision loss. Mark Morad, who lost his sight over three years, launched the tool after less than two years of development. His sister, who is also visually impaired, is among its early users.
The app uses a smartphone camera to capture scenes, then applies computer vision and natural language generation to describe objects, text, and people in real time. Morad said the idea came from his own frustration with existing assistive tools, which he found slow, expensive, or limited to English.
"I wanted something that felt like a friend describing the room to me, not a machine reading labels," Mark Morad, founder.
The system supports Arabic and English, a key advantage in Egypt where most commercial vision apps lack strong Arabic language models. Morad worked with local developers and used open datasets to train the recognition engine. He declined to name the specific model architecture, citing competitive reasons.
From personal need to a tool for thousands #
Morad's sight began deteriorating due to a progressive retinal condition. He said the hardest moment was realizing he could no longer see his sister's face. That pushed him to learn programming and machine learning, despite having no formal background in either.
The app now has thousands of downloads, according to Morad. Users can point the camera at a menu, a medicine bottle, or a street sign and hear a spoken description within seconds. The app also reads handwritten notes, which many mainstream tools struggle to handle in Arabic script.
Assistive technology researchers note that Arabic-language computer vision remains underfunded. A 2025 report from the World Health Organization estimated that over 90% of visually impaired people live in low- and middle-income countries, where localized tools are scarce. Morad's work fills a gap that global tech giants have largely ignored.
A quiet shift in accessible AI #
Morad is not alone in this niche. Several startups in the Middle East and North Africa are now building Arabic-first assistive AI. But most rely on cloud APIs from major vendors, which raises privacy and cost concerns. Morad's app processes some tasks on-device, reducing latency and keeping sensitive data local.
He plans to open-source parts of the Arabic text recognition module next year. That could help other developers build regional tools faster. For now, his focus is on improving scene descriptions for low-light conditions and crowded spaces, two common failure points in current systems.
The broader assistive AI market is projected to surpass $12 billion by 2028, driven by advances in multimodal models and cheaper edge hardware. Morad's story shows that meaningful innovation can come from lived experience, not just corporate labs.