UnifyMed

• 

  Workflow

• 

  Home Page

• 

  Dashboard

• 

  Dashboard

• 

  Unified Medical Report

Inspiration

Medical records are often scattered across countries, languages, and healthcare systems, making it difficult for doctors to get a complete, clear picture of a patient’s history. We wanted to eliminate language barriers and standardize drug names so that physicians can make faster, safer decisions, regardless of where the patient was treated before.

What it does

UnifyMed scans and processes medical reports from any country or language using OCR, translates them into the doctor’s preferred language, and standardizes all drug names using RxNorm CUIs. The result is a unified, structured patient health history that doctors can read at a glance.

How we built it

• For image to text we used OCR technology (Gemini OCR) to extract text from medical reports.
• Translation & Parsing: Gemini API processes the text, extracts key medical terms, and translates them into English (or doctor’s language).
• Drug Standardization: Drug names are sent to the RxNorm API, which returns the official U.S. standard names using RxCUIs.
• Report Generation: The data is auto-filled into a structured patient report template based on doctor's needs.

Challenges we ran into

• Integrating RxNorm was tricky because it only accepts official drug names and sometimes failed on foreign brand names. We had to preprocess input (transliterate, translate, and map common foreign brands to their active ingredients) to get reliable results.

Accomplishments that we’re proud of

• Successfully handled multiple languages including Spanish, Russian, Chinese,French and many more with high accuracy.
• Automatically mapped international drug brand names to U.S. equivalents.
• Built a workflow that gives doctors a single, clear view of a patient’s longitudinal health record.

What we learned

• Handling international medical data isn’t just about translation, drug name standardization is a huge part of interoperability.
• RxNorm is powerful, but you must clean and normalize input before querying for accurate matches.
• OCR quality and preprocessing can make or break downstream analysis.

What’s next for UnifyMed

• Expand to support ICD-10 and SNOMED CT coding for diagnoses and procedures.
• Build a patient portal for uploading records directly and tracking their health history.

Built With

• gemini
• next.js
• ocr
• rxnorm
• translationapi
• vercel
• visionapi