Agentic-AI Driven System for Real-Time Disaster Alerts

This research paper presents an AI-powered Disaster Prediction and Alert System that predicts the probability of natural disasters such as floods, cyclones, heatwaves, and landslides using real-time weather data and machine-learning inference through OpenRouter AI models.The system integrates Flask for backend processing, MongoDB for scalable document-based data storage, and external APIs including OpenWeather, NewsAPI, and IP geolocation services.The model analyzes weather parameters such as temperature, humidity, wind speed, pressure, and rainfall to generate risk levels.Features such as prediction history, automated email reporting through SMTP, fallback news generation, and an admin monitoring panel enhance the overall robustness of the system.Experimental results showed stable performance, accurate predictions aligned with real meteorological behavior, and efficient data retrieval.The study demonstrates the feasibility of combining cloud APIs and AI inference models to create a reliable disaster-awareness platform..