In one embodiment, a method includes storing a caption file including a set of visual representations of captions for a media program and a video file including video information for the media program. A server receives a request for the media program where the request is associated with a user device. The method determines the caption file and the video file for the media program based on the request. The server then sends visual representations of the captions from the caption file out of band with video information from the video file over a network to the user device. The user device is configured to overlay a visual representation of a caption over a portion of the video information.