We introduce Evolving Systems of Stochastic Differential Equations. This model generalizes the well-known stochastic differential equations with Markovian switching, enabling the countably many local systems to have solutions in regime-dependent dimension. We provide two constructions, the first one based upon general results on measure-valued processes and the second one partially inspired by recent developments of the theory of concatenation of right processes. We prove the Feller property under very mild assumptions, provide some extensions to the basic model, and show applications of our general framework to a biological model.