Fyllo launches DeepMet 1, AI weather model for 6-hour rainfall alerts

File Photo: Paddy crop ready to be sown submerged in water in a paddy field at Ariyalur Thirukkai near Villupuram due to heavy rains on Monday. Photo: The Hindu
| Photo Credit:
KUMAR SS

Agritech startup Fyllo has launched DeepMet 1, an AI-powered weather prediction model designed to provide short-term rainfall forecasts of up to six hours with enhanced accuracy for farmers, cities, and industries that rely on precise weather information.

A media statement said DeepMet 1 operates at a hyperlocal resolution of 10×10 km and has been specifically developed for the Indian subcontinent. The model aims to bridge the gap between traditional numerical weather prediction (NWP) systems and the real-world needs of India’s diverse climatic zones. It addresses challenges such as fluctuating crop yields, urban disruptions, and strain on power systems caused by unexpected rainfall.

By analysing satellite observations from the previous six hours, DeepMet 1 produces highly localised precipitation forecasts, often distinguishing rainfall variations even between neighbouring villages, streets, or farmland blocks, the statement added.

Live forecast

DeepMet 1 is designed for real-time deployment and integrates seamlessly via API access, enabling instant adoption across sectors, including agriculture, urban infrastructure planning, energy management, logistics, research institutions, and financial services.

Its forecasts can be publicly visualised on a live dashboard with APIs available for developers and institutions building weather-dependent applications.

Quoting Sumit Sheoran, Co-founder of Fyllo, the statement said: “India’s diverse weather necessitates dynamic tools for accurate forecasting. Farmers, city planners, and energy managers often rely on general forecasts that overlook local conditions. DeepMet 1 provides hyperlocal, AI-driven weather intelligence, making it accessible to all stakeholders.”

Published on November 24, 2025