Megunto
The app version of an existing exam-prep platform for doctors and medical students, with AI-assisted evaluation – for iOS, Android and web.
- Role
- Period
- Stack
- Store rating
Starting point
Megunto existed as a web platform for exam preparation with simulations, evaluations and its own backend. Users study on the go and expected an app. The backend was not to be touched.
What I did
I built the app for iOS, Android and web as a Flutter client on top of the existing backend: login via Auth0, exam simulations with text and voice input (recorded in the app, transcribed via the backend API), questions read aloud and an audio mode for studying without looking at the screen. I implemented the AI-assisted evaluation of learning progress and weak-spot analysis in an interface that keeps interaction as simple as possible. In the web version, courses can be bought via Stripe. Design in Figma, release to both stores.
The API client is generated from the backend’s OpenAPI spec (Retrofit, Dio), so app and backend can’t drift apart. The app is structured into services and controllers (get_it, ChangeNotifier), with routing via auto_route. GitHub Actions builds the web version as a container and rolls it out to Google Cloud Run in three environments.
Result
The app is live in both stores and used in medical institutions, store rating 5/5. The app’s UX concepts were adopted as the blueprint for the relaunch of the web platform.
Stack
Flutter, Dart, get_it, auto_route, Retrofit/Dio (OpenAPI), Auth0, Stripe, GitHub Actions, Google Cloud Run, Figma, App Store Connect, Google Play Console