Skip to content

Plasma

Our innovative AI prescription verification platform transforms how medical prescriptions are validated.

Code Video Cerebral Valley

Core trick

Runs Gemini OCR on prescription images, then fuzzy-matches the doctor's name against the NMC registry.

Tags

healthvision

Description

Our innovative AI prescription verification platform transforms how medical prescriptions are validated. Built with Google Gemini technology, the system automates the entire verification workflow that traditionally required time-consuming phone calls and manual processes from pharmacy services.
Dan-Arnin/Intelligent-Drug-Validator-Cum-Ordering-agent★ 00 forksPythonCreated Feb 14, 2026 · last push Feb 14, 2026

README excerpt

# Medicine Verification Service

A FastAPI-based service for extracting prescription details from medical documents using Google Gemini 3 OCR.

## Features

- 📄 **Multi-format Support**: Upload PDFs, JPG, JPEG, or PNG prescription images
- 🤖 **AI-Powered OCR**: Uses Google Gemini 3 Pro for intelligent text extraction
- 📊 **Structured Data**: Extracts doctor info, patient details, and medicine information
- 👨‍⚕️ **Doctor Verification**: Validates doctor credentials against National Medical Council (NMC) registry
- 🔍 **Intelligent Name Matching**: Uses fuzzy matching to verify doctor names
- ⚡ **Fast & Scalable**: Built with FastAPI for high performance

## Extracted Information

The service extracts:

### Doctor Information
- Hospital Name and Address
- Doctor Name
- Registration Number

### Patient Information
- Patient Name
- Age
- Patient ID
- Prescription Date

### Medicine Details
- Medicine Name
- Dosage (e.g., 500mg, 10ml)
- Dosage Instructions (e.g., 1-0-1, twice daily)
- Timing (AF - After Food, BF - Before Food)
- Duration (e.g., 5 days, 1 week)

## Installation

1. **Install dependencies**:
```bash
pip install -r requirements.txt
```

2. **Set up environment variables**:
Create a `.env` file with:
```
GEMINI_API_KEY=your_gemini_api_key_here
```

## Running the Service

### Development Mode
```bash
python main.py
```

Or using uvicorn directly:
```bash
uvicorn main:app --reload --host 0.0.0.0 --port 8000
```

The service will be available at:
- API: http://localhost:

Team