Features are disclosed for facilitating content presentation and consumption based on input regarding a user's gaze. Gaze tracking or eye tracking information can be obtained from, e.g., a user-facing camera. A content presentation system can determine whether the user's eyes have closed and how long they have closed, whether the user's gaze is directed at the portion of content currently being presented or at some other location, etc. Based on such determinations, presentation of content may be modified. In some cases, a current presentation position may be changed, highlights may be applied, a special bookmark may be generated, etc.