As the energy system is becoming more complex and weather-dependent, weather forecasts are of essential importance for transforming the energy landscape towards decarbonization. In this project, we combine probabilistic numerical weather prediction with machine learning algorithms in order to provide optimized decision-support for green energy production and consumption. Probabilistic forecasts have been underused in this context due to complexity related to their interpretation and data volume. To solve this we will create a machine-readable data portal coupled to machine learning algorithms. We will tailor improved weather forecasts to better support the needs of green energy applications. Our smart solutions optimize specific energy systems based on the latest weather forecast, local observations, and available production from renewable energy. Targeted appliances are heating, ventilation and air-conditioning in buildings, battery charging, and solar and wind power applications.