In one embodiment, a method receiving a request for a media program from a user device. The method then determines a set of visual representations of captions for the media program and determines video information for the media program. Visual representations from the set of visual representations of captions are sent with the video information over a network to the user device where text for the captions has been pre-rendered in the sent visual representations before sending of the visual representations to the user device. Also, the user device is configured to directly render and overlay a visual representation of a caption from the visual representations over a portion of the video information without rendering of the text for caption on the portion of the video information at the user device.