Smart Health Receipts

Inspiration

What it does

Inspiration

Displacement forces you to rebuild your life from scratch. As a refugee currently navigating life's complexities, I realized that managing healthcare and medical expenses during transitions is a chaotic nightmare. You deal with lost paper receipts, fluctuating foreign currencies, and a constant lack of stable internet connectivity. I needed a tool that was as resilient as the people who need it. I was inspired to build a privacy-first, offline-capable medical financial assistant that empowers users to take control of their health expenses, no matter where they are or what currency they are dealing with.

What it does

"Smart Health Receipts" is a 100% client-side, AI-powered medical expense tracker.

• Offline AI Scanning: It uses on-device OCR to read and extract data from medical bills instantly.
• Smart Categorization: Automatically classifies receipts (Hospital, Pharmacy, Lab, Clinic) based on extracted text.
• Real-Time Currency Conversion: Automatically detects foreign currencies (USD, EUR, SAR) on the receipt and converts them to the local currency (LYD) using live exchange rates.
• Dynamic Localization: Fully supports English, French, and Arabic with on-the-fly UI and graph translation.
• Duplicate Prevention: Features a smart text-hashing mechanism to prevent identical receipts from being logged twice.

How we built it

I built the application using React.js. To ensure absolute data privacy and solve the issue of unstable internet connections, I engineered a Local AI Engine using Tesseract.js. Instead of sending sensitive medical data to cloud servers (which can fail or compromise privacy), the app processes the image directly inside the user's browser. I also integrated the Open ER-API to fetch real-time exchange rates, combining it with regular expressions to parse amounts and currencies dynamically.

Challenges we ran into

This is where the project truly tested my limits. I built, debugged, and deployed this entire application exclusively on a 6-inch smartphone screen using Termux. Living through the realities of displacement means dealing with severe infrastructure challenges: frequent power outages, severely limited internet connectivity, and the absence of a proper development machine.

When my initial attempts to use cloud-based LLMs failed due to network timeouts and API rate limits, I couldn't just give up. The constraints forced me to innovate. I pivoted completely to a local-first architecture. Typing complex Regex patterns, managing React state, and configuring Webpack for Tesseract.js on a virtual mobile keyboard wasn't just difficult—it was a crucible that forced me to write ruthless, highly-optimized, and resilient code.

Accomplishments that we're proud of

I am profoundly proud that I didn't let my circumstances dictate my output. Building a sophisticated, multilingual, AI-driven web application entirely on a mobile phone is my biggest accomplishment. I successfully engineered a solution where the AI processing happens entirely on the client-side edge, ensuring zero server costs and 100% medical data privacy for the user.

What we learned

Constraints breed ultimate innovation. I learned that having limited resources (no laptop, no stable internet) forces you to think deeply about performance, offline capabilities, and architectural efficiency. I mastered on-device OCR, dynamic state management in React, and most importantly, I proved to myself that a determined mindset can overcome any hardware limitation.

What's next for Smart Health Receipts

I plan to introduce exportable PDF reports for insurance claims, add predictive analytics to forecast monthly health budgets, and eventually integrate optional encrypted cloud-syncing for users who have stabilized their living situations.

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for Smart Health Receipts

Built With

• artificial-intelligence
• css3
• html5
• javascript
• offline-first
• react
• termux
• tesseract.js