UC Davis is developing a control system that uses model predictive controls to optimize heat pump operation for cost, comfort, and carbon emissions. As part of CalFlexHub, this work is getting expanded to work with real-time electricity price signals. The control works as a supervisor control for heat pump space conditioning systems or water heaters that have an existing API which enables communication with the cloud. The control algorithm sends the on-board control system a new setpoint schedule every 15 minutes based on the objective of satisfying comfort while reducing cost, carbon emissions or both. The control system is being developed as part of a modular control framework (data model framework) to allow two different heat pump systems to be controlled using the same framework and minimal modifications.