@Windystu You are looking at model data that is included in the Windy API. There is always an inaccuracy present in weather models. Some models are better in the boundary layer, other models at other levels of the atmosphere. Some are better with convective precipitation, others with dynamic precipitation. Models differ in resolution, used numerical equations, surface/soil diagrams, hydrostatica etc. etc.
In the end, when looking at a computer generated forecast, you are looking at the output of a single numerical weather model. Some understanding of how the weather and how numerical weather models work is essential in creating a realistic forecast. The fact that models itself are less accurate could be contributed to the synoptic situation at that moment, or changes in the calculations/models. It is not attributed to the API/software you are using while looking at the chosen model.