AnatoMate.ai

✨ Inspiration

Healthcare data organization is broken. Over $1 trillion is wasted annually on healthcare inefficiencies in the U.S., with up to 75% of that waste linked directly to fragmented records and poor coordination.

As international students, when we moved to new countries, our entire medical history got left behind — diagnoses, vaccinations, surgeries — gone. New doctors, new hospitals, but zero access to our own health stories.

We realized this wasn’t just our problem — it’s a global, everyday issue.
So we set out to build a platform that puts patients back in control of their health data, where it belongs.

🩺 What it does

📋 User-Owned Medical Records
Patients store and manage their full medical history directly on their devices. No more fragmented records across hospitals. Ownership is verified using blockchain smart contracts, ensuring integrity and authenticity.
⚙️ On-Demand Data Sharing via QR Code
Need to share your medical history with a new doctor? Just scan a secure QR code. You control what’s shared, when, and with whom.
📊 Health Tracking and Personalized Insights
Track your appointments, prescriptions, and vitals in one place. Our platform uses your data (only with your permission) to provide personalized health summaries.
🧠 AI-Powered Health Risk Prediction
Gradient boosting and deep learning models analyze longitudinal health data to predict risk factors — proactively helping users take control of their future health outcomes.
🔗 Blockchain Integrity Layer
Every record update is hashed and anchored immutably onto the Polygon blockchain, ensuring tamper-proof medical proof without exposing private data.

🛠️ How we built it

Frontend

Built with Next.js, Tailwind CSS, and React component libraries to ensure a fast, responsive UI.

Blockchain Integration

Smart contracts written in Solidity, deployed on Polygon PoS (for low fees and fast transactions).
Frontend blockchain interaction handled with ethers.js and wagmi for seamless wallet connections.

Backend Services

Built a lightweight backend using FastAPI (Python) to handle API calls, authentication handshakes, and data aggregation securely.
MongoDB was used to store non-sensitive metadata such as user settings, appointment histories, and AI model outputs.

Authentication

Implemented secure, OAuth 2.0-compliant user authentication and authorization using Auth0, making login and access management seamless across devices.

QR Code Data Sharing

Generated scannable, encrypted QR codes using qrcode.react and react-qr-reader to facilitate secure, on-demand record sharing without NFC.

AI & ML

Trained gradient boosting and deep learning models using scikit-learn and TensorFlow on synthetic healthcare datasets to predict personalized risk factors based on historical trends.

3D Data Visualization

Used Three.js to create immersive and interactive 3D visualizations of the patient’s medical history and vitals over time, making complex information more intuitive and accessible.

🧩 Challenges we ran into

We were ambitious — maybe too ambitious.
We set out to solve a problem spanning blockchain, AI, QR-based secure sharing, user-owned data, and privacy-first healthcare — all in just 36 hours.

At one point, it felt like the entire project was collapsing under its own weight.

Technical Complexity Explosion
We originally envisioned NFC-based peer-to-peer data sharing, but halfway through we discovered that iOS devices don’t allow peer-to-peer NFC.
That forced a complete pivot to QR code sharing, requiring us to rethink encryption, scanning UX, and data handoff flow in a matter of hours.
AI Model Training under Fire
Training machine learning models to predict health risks sounded exciting — until we realized real patient datasets are restricted for privacy.
We scrambled to generate synthetic healthcare datasets just to have something realistic to work with. Creating a pipeline that was believable, yet compliant, while the clock was ticking, pushed us to our limits.
Backend–Frontend–Blockchain Coordination
Coordinating between FastAPI backend services, MongoDB storage, Next.js frontend, and Polygon blockchain interaction — all while integrating Auth0 for authentication and Three.js for visualization —
felt like conducting an orchestra where every musician was playing in a different key for the first few hours.
Privacy vs Blockchain Tension
We knew we wanted to use blockchain for proof of record authenticity, but we also knew we couldn’t expose patient health data on-chain.
It took multiple late-night whiteboard sessions (and a few failed smart contract drafts) before we finally arrived at the right balance: hash anchoring without storing raw data.

Despite all this, we adapted.
We cut ruthlessly where we had to.
We made bold technical pivots.
We held the vision steady even when the architecture kept shifting.

By the end, we didn’t just build what we set out to — we built something even stronger.

🏆 Accomplishments that we're proud of

Built a fully working prototype in just 36 hours — including blockchain integration, QR code sharing, and AI modeling.
Designed a secure, privacy-preserving architecture that respects users’ rights to their own data.
Pushed ourselves across domains — blockchain, healthcare, AI, and frontend — and made them work together in a real, usable product.