A network-based, e.g., cloud-based, and apparatus based in-call command control for dynamically performing user commands such as voice commands or gesture commands are provided. The result allows for hands free operation of a calling system including one where the user is located separate from a base station, and provides for recognizing both voice and gestural commands. Filtering of commands can be performed dynamically so that commands are not perceived by those participating in the call. Further, the user, who has given the command, can be notified regarding the performance of the requested command.