Recommending and activating scenes in a multimedia system are provided. The techniques of the present disclosure can allow the states of the components in the multimedia system to be captured in a scene. Once the scene has been saved, the scene can be activated at a later time. A controller device for recommending scenes by comparing the current state with states of components in saved scenes is also provided. The controller device can also determine if media is playing in a current scene, and activate a selected scene without playing the media content. Once the selected scene is activated, the media playing in the current scene can be switched, and the media playing in the selected scene can be played.