Inspiration 💡
The inspiration for ClarityMD came from a very personal experience. While living as an expat in Vietnam, I faced health issues that required multiple doctor visits. Every appointment ended with pages of medical results — in Vietnamese, full of medical jargon, and difficult to interpret even after translation.
The language barrier combined with the complexity of medical terminology left me frustrated, anxious, and unsure about my health. That’s when the idea sparked: “What if AI could bridge the gap between confusing medical results and clear, patient-friendly understanding — no matter the language?”
ClarityMD was born out of that need — to make medical knowledge more accessible, compassionate, and empowering for patients worldwide.
What We Learned 📚
How to optimize Groq’s GPT-OSS for real-time (<3s) medical text simplification.
The importance of UX in healthcare apps: calming colors, trust indicators, and readability adjustments make a difference.
How to handle OCR and multilingual workflows (English, Spanish, French, German, Italian, Portuguese, Vietnamese).
Practical lessons about security & privacy in health-related software (rate limiting, no data storage, secure API handling).
How We Built It 🛠️
Primarily using replit.
Frontend
Built with HTML5, CSS3, and Vanilla JavaScript. Added drag-and-drop uploads, responsive UI, accordion dictionaries, and download features. Designed a landing page + app page for a polished judge experience.
Backend
Flask (Python) server with /process endpoint. Integrated Groq GPT-OSS-120b for medical simplification. Returned structured JSON: simplified text, highlighted terms, and suggested doctor questions.
OCR + Translation Pipeline Tesseract.js for client-side OCR (privacy-first). Auto-detection + translation of documents before AI analysis.
Performance & Security
Single API call optimization for speed. avg <3s response times showcased with a real-time timer. Secure environment handling for API keys.
Challenges We Ran Into 🚧
Balancing simplicity and accuracy: making content easy while preserving critical medical details. Multilingual OCR quirks: noisy scans required extra preprocessing. Time pressure: optimizing Groq’s large model responses in real time. Designing for trust: building UI elements that feel professional and reliable in a sensitive domain.
For the Judges ✨
ClarityMD isn’t just a demo — it’s a vision for patient empowerment. By combining AI speed, structured outputs, and accessibility-first design, it has the potential to transform healthcare communication globally.
Log in or sign up for Devpost to join the conversation.