GlucoGuard AI

GlucoGuard AI is a fully offline, local-first clinical decision support system designed specifically to monitor, assess, and simulate patient progression towards Type-2 Diabetes using mathematical baselines and Machine Learning counterfactuals.

Features Included

• Monorepo Architecture: Python FastAPI backend & React Vite frontend seamlessly running side-by-side.
• Offline Backend & DB: Uses a local SQLite database.db file instead of Supabase, ensuring full offline capability for clinician privacy.
• Role-Based Views: Optimized for clinician-only workflow with deep insights and specialized SHAP feature explainability.
• Risk Simulator: A powerful UI that allows clinicians to run "What if?" counterfactual simulations, querying the AI to observe absolute risk reduction.
• Bulk Import Data: Add massive patient rosters effortlessly via CSV uploads directly in the Patient View.

Installation & Setup

Ensure you have Python 3.x and Node.js 18+ installed.

1-Click Start

To boot up both the Web Frontend and the FastAPI Backend with a single click, just run the batch file:

./start.bat

(This will launch the services in background and tail logs. Closing the script will terminate the servers)

Manual Setup

If you prefer starting them individually:

Terminal 1 (Backend):

cd server
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

This automatically initiates the SQLite database locally at server/glucoguard.db.

Terminal 2 (Frontend):

cd client
npm install
npm run dev

Security & Privacy Note

This product is strictly For Clinical Decision Support Only and cannot act as a substitute for professional medical judgment. Because patient data (PHI) is processed entirely offline, data stays directly on the clinician's localized network, requiring no cloud syncs.