Embodiments include method, systems and computer program products for aligning visual content to search term queries. In some embodiments, one or more search terms may be received. A plurality of results may be obtained using the one or more search terms on metadata associated with preprocessed visual media. A subset of frames may be selected for each of the plurality of results. A search results set may be generated by optimizing the subset of frames for each of the plurality of results. The search results set may be transmitted.