The objective of the SHAPED project is to define a software architecture and set of algorithms enabling the most effective processing of speech between the embedded device and the cloud, balancing user experience and operation costs across the range of Logitech voice-enabled interface devices.