I present an unobtrusive multimodal interface for smart objects (MISO) in an everyday indoor environment. MISO uses

pointing for object selection and context-sensitive arm gestures for object control. Finger snaps are used to confirm object selections and to aid with gesture segmentation. Audio feedback is provided during the interaction. The use of a Kinect depth camera allows for a compact system and robustness in varying environments and lighting conditions at low cost.