Inspiration
Watching a friend's grandmother struggle with daily medication management and lose her independence to reliance on others inspired MedVoice. No one should have to compromise their autonomy just to read a prescription label safely, which sparked our mission to build a barrier-free digital assistant.
What it does
MedVoice turns any smartphone camera into a talking pharmacist for blind, low-vision, and hard-of-seeing users. By combining real-time OCR label scanning, slow-paced automated voice readouts, and universal physical controls (where any keystroke triggers the next step), MedVoice ensures patients can independently scan prescription bottles, verify dosages, and track their daily schedules safely.
How we built it
The app is built as a single, ultra-fast web application using HTML5, Tailwind CSS (via CDN), and pure JavaScript (ES6+). It leverages native browser APIs:
navigator.mediaDevices.getUserMedia for live camera vision feed and bounding reticles.
window.speechSynthesis with optimized slow-paced pacing (rate: 0.75) for comprehensive accessibility readouts.
Custom Web Audio API sound synthesizers for tactile auditory feedback.
Challenges we ran into
Designing an interface that remains fully usable without relying strictly on eyesight or precise touchscreen taps was a major design challenge. We solved this by implementing a Universal Key Control system—allowing users to press any physical key on their keyboard to advance prompts or trigger actions—alongside automatic audio indexing for all screen options.
Accomplishments that we're proud of
Successfully engineering a zero-dependency, single-file accessible web app that runs instantly in any browser.
Designing an empathetic user experience specifically tuned for blind and elderly users with high-contrast elements, tactile Peter Tarka-inspired sculptural cards, and deliberate audio pacing.
Translating a deeply personal inspiration story into a functional, life-enhancing prototype.
What we learned
We gained profound insights into inclusive design principles, learning how critical auditory cues, predictable menu structures, and robust keyboard navigation are for making healthcare technology genuinely accessible to everyone.
What's next for MedVoice
Integrating advanced machine learning OCR models (such as Tesseract.js or cloud vision APIs) for offline prescription parsing.
Adding automated push notifications and caregiver alert summaries.
Expanding multi-language voice support to serve diverse communities globally.
Log in or sign up for Devpost to join the conversation.