ReliefNet- Intelligent Disaster Response System

🚨 Offline-First Agentic AI Disaster Response System

🧠 Overview ReliefNetAI is a government-centric agentic AI system that enables citizens to send distress signals (panic, injury, location) and receive verified alerts—even in low or no connectivity environments. It intelligently routes information between citizens and authorities to accelerate emergency response.

This project is a government-centric agentic AI system designed to assist citizens during disasters.

• Government broadcasts alerts and guidance

• Citizens can send distress signals (panic, injury, location)

• Works in hybrid mode:

  • 🌐 Online → Full functionality
  • 📴 Offline → SMS-based communication

🎯 Features

• 🚨 Panic Mode (one-click emergency alert)
• 📍 Location-based assistance
• 📡 SMS fallback (no internet required)
• 🤖 Agentic AI decision-making
• 📚 Offline data support (RAG-like system)
• 🏛️ Government-side monitoring dashboard

🏗️ Tech Stack

• Backend: Python (Flask / FastAPI)
• AI: OpenAI API
• Database: SQLite / JSON
• SMS: Twilio / Device SMS
• Maps: Google Maps API
• Frontend: Streamlit

📲 How It Works

🟢 Online Mode

• User inputs query or presses panic button
• Data sent to server via API
• AI processes and responds

---

🔴 Offline Mode

• App generates SMS with:

  • Location
  • Injury status

• SMS sent to government number

• Server parses SMS and processes it

📡 SMS Format

ID:123 LOC:16.506,80.648 STATUS:CRITICAL MSG:TRAPPED

🎬 Demo Scenario

1. Simulate disaster (e.g., flood)

2. User presses panic button

3. System:

   • Sends location
   • Classifies urgency
   • Suggests rescue actions

🏆 Innovation

• Offline-first AI system
• Hybrid communication (API + SMS)
• Government-centric architecture
• Real-world applicability

🚀 Future Improvements

• Voice input support
• Advanced routing
• Integration with government systems

👨‍💻 Authors

Hackstreet Boys