Agritech startup Fyllo has announced the launch of DeepMet 1, an artificial intelligence-based weather prediction system designed to provide short-term rainfall forecasts of up to six hours. The model focuses on improving weather accuracy for users who depend on precise local data, including farmers, city authorities and industrial players. DeepMet 1 is positioned as a response to the growing demand for more reliable, area-specific climate insights in India.

According to the company, DeepMet 1 offers predictions at a hyperlocal scale, functioning at a resolution of ten by ten kilometres across the Indian subcontinent. This approach attempts to address limitations of traditional numerical weather prediction systems, which often provide broader regional forecasts that overlook rapid and localised changes. The system has been built specifically to reflect India’s diverse geographical and climatic conditions.

Fyllo claims that DeepMet 1 aims to support better planning across agriculture, energy distribution and urban management. It targets climate-linked challenges such as unpredictable rainfall impacting crop yields, traffic disruptions due to sudden showers and weather-related strain on power infrastructure. The company states that its model is built to close the gap between large-scale forecasts and real-time local needs.

Short-Term Forecasting

The firm stated that DeepMet 1 analyses satellite data from the previous six hours to generate localised precipitation predictions. Fyllo says the system can differentiate rainfall patterns between closely located areas, including neighbouring villages, roads or agricultural fields. This feature attempts to refine decision-making for industries heavily influenced by rapidly changing weather.

The agritech company believes India’s climate conditions require dynamic forecasting tools capable of addressing local variations. Farmers, utility managers and city planners often rely on generalised forecasts that may not reflect immediate risks or beneficial weather events. The model has been developed to make AI-based forecasting more accessible for on-ground decision makers rather than only institutional users.

Sumit Sheoran, cofounder of Fyllo, said the initiative is aimed at delivering actionable weather insights to Indian stakeholders who face frequent weather uncertainties. According to Sheoran, hyperlocal intelligence is expected to become an important element of climate resilience as changing weather patterns increasingly affect urban and rural systems.

Company Operations

Fyllo operates as a machine learning-enabled IoT platform within the agricultural sector, offering technology-backed solutions to boost farm productivity. According to the company’s release, it has raised approximately USD 6.52 million from investors, including India Quotient and SIDBI Venture Capital. The organisation continues to expand its technological offerings as digital solutions gain relevance across India’s agriculture ecosystem.

The introduction of DeepMet 1 aligns with new data-driven models emerging in India’s agricultural technologies market. As crop planning, irrigation management and pest control increasingly rely on micro-level climate information, accurate forecasting is becoming an operational necessity rather than a supplementary tool. Fyllo’s platform positions itself within this transition, aiming to bring advanced forecasting to a wider user base.

Agritech Startups in India

India’s agritech sector has been witnessing rapid adoption of artificial intelligence and IoT-driven systems. Startups across the country are building solutions around data analytics, precision farming, climate intelligence and farm automation. As unpredictable weather continues to influence yields and rural incomes, agritech companies are prioritising hyperlocal insights to support farmers with real-time decision tools. The growing focus on weather-based intelligence is expected to expand as climate variability rises, prompting the need for more advanced forecasting technologies in agriculture and allied industries.