MedBuddy-AI

Inspiration

Accessing quick, reliable, and understandable preliminary healthcare advice is a major challenge for many patients. We built MedBuddy-AI to act as an accessible digital medical companion that simplifies medical jargon, evaluates symptoms logically, and provides instant health guidance.

What it does

MedBuddy-AI provides an intuitive interface for real-time healthcare support:

Symptom & Triage Evaluator: Analyzes user-reported symptoms to assess case urgency and provides immediate home-care or emergency recommendations.
Prescription & Medication Guidance: Explains complex drug dosages, usage instructions, and potential side effects in simple terms.
Interactive Health Assistant: A responsive web application allowing users to seamlessly discuss health concerns and get instant AI-driven answers.

How we built it

Backend: Python (Flask/FastAPI) handling API requests, business logic, and prompt orchestration via app.py.
Frontend: Clean, responsive Vanilla HTML, CSS, and JavaScript (index.html, script.js) for fast page loads and zero bundle overhead.
Deployment & Setup: Configured with Procfile and runtime.txt for continuous deployment and environment management.
Devin Integration: Utilized Devin AI to optimize API routing in app.py, refactor JavaScript state handling in script.js, and generate deployment readiness scripts.

Challenges we ran into

Managing lightweight frontend-to-backend API communication without heavy framework overhead.
Ensuring the Python backend handles edge-case inputs gracefully while maintaining low API latency.
Crafting system prompts to keep medical outputs safe, accurate, and non-hallucinatory.

Accomplishments that we're proud of

Built a lightweight, fast, and fully functional healthcare application from scratch.
Created a seamless integration between a Python backend and a zero-dependency frontend.
Ensured high-speed responses while maintaining clean, maintainable code architecture.

What's next for MedBuddy-AI

OCR Report Parsing: Allow users to upload lab report images and parse them instantly.
Multi-Language Support: Expand localized healthcare responses for non-English speakers.

Built With

agentic
ai
api
css3
devin
flask
heroku
html5
javascript
python
rest

Submitted to

Agentic Hackathon

Created by

What I Worked On & Contributed:
Architecture & Backend: Designed and implemented the core multi-agent orchestration logic using Gemini API and Google Cloud infrastructure.
Feature Development: Integrated key health-tech functionalities, including real-time AI guidance, automated symptom evaluation, and dynamic data processing pipelines.
Problem Solving & Refinement: Optimized prompt engineering and API interactions to ensure fast, reliable, and contextual outputs.
What I Learned & What I'm Proud Of:
Successfully built a functional end-to-end AI agent system solving real-world healthcare accessibility problems.
Gained deep, hands-on experience with Gemini models and Google Cloud Agent Builder integrations.
Successfully translated complex user requirements into a seamless UI/UX experience for digital health solutions.

MD. ABDUR RAHMAN HRIDOY
Tech enthusiast and college student. Eager to learn, collaborate, and build innovative projects at hackathons.
MD Riad Hossain