MediSure

Precision in Every Prescription

Comment

Inspiration

Access to reliable medical information is often scattered, hard to interpret, and not easily available to patients or healthcare providers in real time. We wanted to create a solution that leverages AI to make understanding prescriptions, drug interactions, and medical information simple, fast, and accessible to everyone. This became the motivation for building MediSure.

What it does

MediSure is an AI-powered medical assistant that:

  • Extracts text from prescriptions using OCR (Tesseract).
  • Provides drug information, usage details, and side effects.
  • Suggests safe drug combinations and highlights potential conflicts.
  • Acts as a quick reference for patients and healthcare professionals to ensure safer decision-making.

How we built it

  • Frontend: Built with HTML, CSS, Javascript for a clean and responsive UI.
  • Backend: Flask handling OCR processing and AI interactions.
  • OCR: Tesseract OCR for extracting text from scanned prescriptions.
  • AI/LLM: Integrated Llama model to provide explanations and drug interaction checks.
  • Database: MongoDB for storing user queries and medical references.
  • Hosting: Deployed on cloud (Heroku/Render/Vercel).

Challenges we ran into

  • Getting accurate text extraction from low-quality prescription images.
  • Handling medical terminology and ensuring AI-generated responses were reliable.
  • Integrating Tesseract properly across different environments (system vs. virtualenv).
  • Building trust in the AI outputs by cross-verifying with medical datasets.

Accomplishments that we're proud of

  • Successfully built an end-to-end pipeline from prescription image → OCR → AI medical insights.
  • Designed an intuitive interface that makes medical information accessible even to non-technical users.
  • Overcame technical hurdles with OCR installation and cross-platform compatibility.
  • Created a project with direct real-world medical relevance.

What we learned

  • Importance of handling unstructured healthcare data with care.
  • How to combine OCR, AI, and medical knowledge bases into one solution.
  • Deployment challenges with dependencies like Tesseract in cloud environments.
  • The need for human oversight in medical AI systems.

What's next for MediSure

  • Expanding the drug interaction database with verified medical sources.
  • Adding multilingual support for prescriptions in different languages.
  • Improving OCR accuracy using deep learning models for handwriting.
  • Partnering with healthcare providers to test MediSure in real-world settings.
  • Adding secure login and history tracking for users to store and revisit queries.

Built With

Share this project:

Updates