Truveo uses a three-pronged search and indexing process: Its crawler searches the transcript text of a video and the text of Web pages that accompany a video, and also recognizes the visual characteristics of a Web application with video, said co-founder and CEO Tim Tuttle.
"Our technology basically takes a screen shot of a Web page and is able to examine the entire object space to determine whether it contains video or not," Tuttle explained. "There's no need for text. Tuttle added that the Truveo doesn't need to rely on content publishers embedding keywords, or meta-tags, on video clips.
Google scans metatags video clips, while Yahoo! searches for keywords and media files. The privately held Blinkx also uses voice recognition to match with search terms.
Like most video search engines, Truveo points users to third-party sites for viewing, in hopes of avoiding licensing issues.
Truveo has thus far indexed millions of feeds, said Tuttle, and expects to significantly improve the breadth and quality of its search listings over the next six months.
Truveo currently is in discussions with a number of publishers who have expressed interest in his search technology, Tuttle added.