Adaptive reading and learning app for an edtech startup

Around 1 in 10 people has dyslexia - the most common learning difficulty in the world. That's hundreds of millions of readers who meet the same text everyone else does, but pay far more to get through it.

The cost isn't just speed. When reading takes real effort, people tire faster, retain less, and often give up before the content ever lands. And this has nothing to do with intelligence or motivation - the words simply arrive in a format that works against them.

Omoguru set out to change that format. Their mission is simple: build concrete, useful reading tools that make life easier and give people back the pleasure of reading. Lexie is where that mission became an app.

Most digital reading assumes one kind of reader - fast, comfortable with dense text, unbothered by standard typography. For a large minority, that assumption quietly fails.

So the goal wasn't a nicer e-reader. It was a reading environment that bends to the person, not the other way around. That meant getting several hard things right at once: typography tuned for readability, visual cues that reduce common mistakes, and audio that stays perfectly in step with the text on screen.

Each piece is subtle on its own. Together, they decide whether someone finishes a chapter or closes the app.

We built Lexie as a progressive web app so it runs on desktops, tablets, and phones from a single codebase - no install friction, no separate builds to maintain.

Inside it, readers shape the text to suit them. They can switch to specialized fonts designed for easier reading, and turn on color coding for confusable letter pairs like b/d and p/q, so shapes that trip people up become easy to tell apart. A line highlight guides the eye across the page and keeps a reader from losing their place in a block of text.

None of this asks the reader to learn a new tool. The controls stay simple; the adaptation happens underneath.

Some readers move through text far more comfortably by listening. So Lexie reads aloud - and keeps the eye and the ear together.

Using Azure Cognitive Services for text-to-speech, the app narrates the content while a focus line follows along word by word. Reading and hearing stay in sync, which is exactly what helps the meaning stick. It turns passive listening into active reading, and it lets people switch between the two whenever their focus shifts.

A reading tool only matters if it works on the things people actually need to read. So Lexie lets users bring in their own documents and ebooks and turn them into material they can read comfortably or listen to.

That's the difference between a demo and a daily habit. School texts, work documents, a book someone's been meaning to finish - all of it becomes approachable in the same adapted environment.

Under the hood, Lexie runs on .NET, Angular, TypeScript, and PostgreSQL, with Azure Cognitive Services handling speech. It's a stack chosen for predictable behavior and long-term maintainability - the kind of foundation that lets a product keep adding features without accumulating fragility.

That foundation paid off. Lexie shipped to production in 2021 and has since grown into a broader reading and learning platform, built on the same core it started with.

Lexie takes the same text everyone else reads and makes it work for the people it usually leaves behind.

The result is reading that costs less effort, learning that survives the first attempt, and a tool people return to because it fits how they actually read. Most importantly, it reframes reading difficulty as a design problem - one that software can solve, right where it matters, on the reader's own screen.