Provided herein are systems and methods of inserting content into videos based on associated text. A video content server can receive a request for video content into a video content slot of a video item played on the client device. The request can be generated responsive to execution of an encoding embedded in the video item. The video content server can identify words derived from a segment of the video item playable prior to the video content slot. The video content server can determine a topical category for the segment slot based on the identified words. The video content server can select a secondary video content item based on the topical category of the segment of the video item. The video content server can provide the secondary video content item to the client device to insert into the video content slot during the video item played on the client device.