Most weather services around the world prepare their forecasts based on a single global NWP model and then apply a local model such as WRF to the data. Meteosource uses a number of data sources and proprietary numerical weather prediction models to extract the most valuable information from each. It learns from historical data and forecasts from many different models, compares their data from different places and meteorological situations, and creates a single output that should minimise inaccuracies and biases of individual models.

API documentation available at https://www.meteosource.com/documentation

I greatly appreciate any feedback and thoughts on how to improve this service and make it easier to use!