Back to projects

Prescription AI Webapp

Decodes handwritten prescriptions into clear text using a vision model + MERN stack.

Vision model extractionReadable post-processingUser-friendly workflow

Problem

Handwritten prescriptions are hard to read and error-prone.

Approach

Used a vision-capable model with a web workflow for clean, readable output.

Value

Improves clarity and reduces errors for patients and pharmacies.

Snapshot

An end-to-end workflow that turns scanned prescriptions into clean, readable summaries with verification steps.

Stack

  • MERN
  • Gemini Vision
  • Post-processing

Role

  • Full-stack build
  • Model integration
  • UX flow

Outcomes

  • Readable output
  • Lower error risk
  • End-to-end workflow

Build notes

  • Vision extraction with a validation layer.
  • Post-processing rules to normalize medicine names.
  • Review UI for fast approval and export.

Roadmap

  • Confidence scoring per line item.
  • Pharmacy-ready export formats.
  • Multi-language support.

Want something similar built for your product?

I'll scope the path, then ship with reliability in mind.