💊 MedWise: AI-Powered Medication Affordability Platform

💡 Inspiration

In the U.S. alone, millions of people struggle to afford their prescriptions—often unaware that generic alternatives exist at a fraction of the cost. The healthcare system is flooded with fragmented drug information, making it hard for patients to compare options. We created MedWise to simplify this process using AI and structured data, helping users make smarter, cost-effective medication decisions with just a search—or a photo of their prescription.


🔍 What It Actually Is

MedWise is a smart medication dashboard that connects users to cheaper, clinically equivalent alternatives to expensive brand-name drugs. It combines a powerful knowledge graph with AI to surface personalized, affordable options in two ways:

1. Manual Search Mode

Search any brand-name drug to view:

  • Generic equivalents
  • Prices
  • Active ingredients
  • Side effects
  • Drug interactions & treated conditions
  • Manufacturer details

2. Prescription Scan Mode (New)

Upload a prescription image and let AI handle the rest:

  • 🧠 Gemini Vision parses drug names from the photo
  • 🔗 Queries our Neo4j GraphRAG pipeline
  • 💡 Returns cheaper, therapeutically equivalent medications automatically

🛠️ How We Built It

  • Frontend: Next.js, React, Tailwind CSS, TypeScript
  • Backend: Flask API connected to Neo4j AuraDB
  • Graph: 135+ nodes and 140+ relationships including prices, manufacturers, ingredients, and medical uses
  • AI Integration:
    • Gemini Vision model for prescription parsing
    • Hugging Face + Neo4j GraphRAG for medication reasoning
  • Performance: Redis caching, API routing via Next.js → Flask backend

⚔️ Challenges

  • Handling variability in prescription formats (digital, printed, handwritten)
  • Aligning extracted drug names with structured graph entities
  • Designing an intuitive UX that makes complex medical data approachable

✅ Accomplishments

  • Functional image-to-alternative medication pipeline
  • Accurate drug parsing using Gemini Vision
  • Built a scalable, visually interactive medication graph
  • Helped users explore and understand how to save on prescriptions

📚 What We Learned

  • How to combine computer vision and knowledge graphs for a meaningful healthcare solution
  • How GraphRAG enhances natural drug search by reasoning over relationships, not just keywords
  • The importance of user-friendly interfaces in medical data systems

🔮 What’s Next

  • Better support for handwritten prescriptions and blurry scans
  • Multilingual expansion for non-English prescriptions
  • Adding pharmacy APIs to show live pricing and availability
  • User accounts to track history and save favorite medications

Built With

Share this project:

Updates