Techniques are disclosed for performing a computer-implemented processing of slide presentation videos to automatically generate index locations corresponding to particular slides within a slide presentation video. In embodiments, a slide presentation video is uploaded to a video processing system. The video processing system performs an image analysis to identify each slide within the slide presentation and determine a time window for each occurrence of each slide. An audio analysis is performed to adjust the time window to the start of a sentence that precedes the introduction of the slide. A user interface includes one or more selectable links associated with each slide that link to a corresponding location within the slide presentation video. Similarly, a processed slide presentation video includes selectable links to index to the corresponding slide of the presentation.