Abstract: Methods and apparatus, including computer program products, implementing and using techniques for converting a target file into a multimedia presentation. A target file including text is received. A class of the target file is identified based on characteristics of the target file. A multimedia presentation is created that contains speech rendered from a portion of the text, additional text, wherein the portion of the text is selected in accordance with a template programmed with content and logic for the identified class, a visual representation of the document, and one or more manipulations of the visual representation defined in accordance with the content and logic of the template. In response to a user command to run the multimedia presentation, the speech, visual representation, and/or the manipulations are output in accordance with the logic of the template.