weathergpt

byPythagoras

Build WeatherGPT, an SIH 2026 disaster-management web/PWA product that converts local weather data into actionable, role-specific decisions for Farmers, Fishermen and Communities. Create a premium, cinematic, minimal dark-teal interface with animated weather scenes, smooth page transitions, micro-interactions, responsive mobile/desktop design and accessible UX. Include separate pages: Landing, Weather Overview, AI Assistant, 24h/7d Forecast, Multi-Hazard Watch, Alerts and Alert Settings. Implement location search + geolocation, live weather, hourly/7-day forecasts, rainfall/wind/heat/fog/thunderstorm risk detection, animated hazard visualization, role-based recommendations, multilingual-ready UI (English, Hindi, Tamil, Telugu, Bengali, Marathi), grounded conversational AI, voice input/read-aloud, preparedness checklist, offline PWA support and SMS/WhatsApp alert configuration. The AI must answer questions using actual loaded weather data and never invent values; safety-critical answers must recommend official confirmation. Make every navigation link, button, form, CTA, filter and interaction functional. Use realistic API/mock data when live APIs are unavailable, clearly label demo/offline data, and never fake successful alerts or government integrations. Architect the app so IMD/MoES, Bhashini, Open-Meteo, LLM and Twilio integrations can be connected later. Add the business layer: WeatherGPT ClimateShield — organizations can monitor farms/assets, receive weather-risk scores, get AI-recommended actions, automate response playbooks and eventually integrate climate-risk/parametric insurance. Demonstrate an agriculture use case where WeatherGPT detects weather risk → identifies affected locations → recommends an action → sends a simulated multilingual alert. Make the final product feel like a real startup-grade climate-risk intelligence platform, not a generic weather app. No placeholder pages, dead buttons, fake dashboards or unnecessary radar UI.