Care AI Vision

AI-powered CARE App (plugin) to extract and auto-fill patient registration details from scanned or photographed forms.

Features

• 📷 Scan or upload — pick an image from camera, gallery, or file picker
• 🤖 AI-powered extraction — uses Care AI backend to extract structured patient data from form images
• 🏛️ Governance resolution — automatically resolves State → District → Local Body → Ward hierarchy via Care's Organization API
• ✅ Auto-fills form fields — name, phone, gender, DOB/age, blood group, address, pincode, and governance location
• 🔁 Retry on failure — reprocesses the same image without re-upload

Getting Started

Prerequisites

• A running Care frontend instance to host this plugin
• A running Care backend with care_ai plugin installed

Installation

git clone https://github.com/ohcnetwork/care_ai_vision.git
cd care_ai_vision
npm install

Configuration

No additional configuration needed! The plugin uses the Care AI backend which is configured at the server level.

The Care AI backend handles all AI provider credentials (Gemini, OpenAI, Azure, etc.) securely on the server side.

Development

npm start

This starts the dev server on port 10120 with hot reload.

Production Build

npm run build

How It Works

1. User clicks "Scan Registration Form" on the patient registration page
2. Browser shows native image picker (camera / gallery / files)
3. Image is sent to Care AI backend (/api/care_ai/ask/) for structured data extraction
4. Care AI backend processes the request using configured AI provider (Gemini, OpenAI, etc.)
5. Extracted fields are validated and auto-filled into the form
6. Governance hierarchy (state/district/local body/ward) is resolved via Care's Organization API
7. User reviews extracted data and confirms