Systems and methods of avatar configuration based on a media stream context are disclosed. In a particular embodiment, a method is disclosed that includes determining context information related to a portion of a media stream. The method also includes selecting configuration settings of an avatar based at least partially on the context information. The avatar is responsive to user input to enable interaction with one or more other users with respect to the media stream. The method further includes sending display data to a user device. The display data includes information to display the avatar with the portion of the media stream.