Agentic Medical Analyser

Agentic Medical Analyser is an AI-powered medical triage and diagnostic assistant designed to simulate intelligent patient intake, risk assessment, department routing, and explainable predictions.

The system combines rule-based logic, optional machine learning models, and LLM-powered assistance to provide fast preliminary medical guidance.

---

Live Demo

https://agentic-medical-analyser.vercel.app

---

Features

Patient Intake

Interactive multi-step form for capturing symptoms and patient details.

• Calls /triage for risk assessment
• Calls /predict for department routing

Triage Results

• Risk level classification
• Suggested medical department
• Confidence score
• Actionable recommendations

AI Assistant

• LLM-powered chat via /chat (Groq API)
• Graceful fallback to local responses if unavailable

AI Explainability

• /explain endpoint provides prediction rationale
• Helps users understand system decisions

Nearby Hospitals

• Location-aware hospital lookup
• Uses /nearest-hospital (Overpass API)

---

Running the Application

The project consists of two components:

• Backend (FastAPI)
• Frontend (Vite / npm)

---

Backend Setup

Run from the Pragyan project root:

cd /Users/apple/Pragyan
python main_combined.py

Default Backend URL: http://localhost:8010

If port 8010 is occupied, the server may automatically use ports 8011–8015.

Backend Requirements

• Python 3.x
• FastAPI
• Uvicorn

Install dependencies if needed:

pip install fastapi uvicorn python-dotenv

Environment Variables

Create a .env file:

GROQ_API_KEY=your_api_key_here

Required for:

• /chat
• /explain

Optional Machine Learning Models

If present, ML-based triage logic will be used:

• models/triage_model.pkl
• models/encoders.pkl

If missing, the system falls back to rule-based logic.

---

Frontend Setup

cd Agentic-Medical-Analyser
npm install
npm run dev

Default Frontend URL: http://localhost:5173

API Configuration

The frontend defaults to http://localhost:8010.

To override:

cp .env.example .env

Edit .env:

VITE_API_URL=http://localhost:8010

---

Conceptual Architecture

User Input → Backend Intelligence Layer →

Component	Endpoint
Risk Engine	/triage
Prediction Engine	/predict
Explainability Engine	/explain
AI Assistant	/chat
Location Services	/nearest-hospital

---

Disclaimer

This project is a research and educational prototype. It is not a replacement for professional medical advice and is not approved for clinical use. Always consult qualified medical professionals for real health concerns.