If you are looking for video content, this is the place to go.
Founded two years ago by a seriously smart team of search experts, Truveo is tackling the very difficult problem of creating metadata from video out of whole cloth – unlike text-based content, it is very difficult to determine context of video and audio content without a human to actually view it (which is error prone and doesn’t scale).
Truveo takes an innvoative two step approach to indexing new content. The first step is a straight web crawl to find the videos. The second, more important step, is to create metadata about the content, beyond what is available in transcripts and feeds (most video content has neither). They’ve created a “visual crawler” that looks at surrounding content to determine context.
Finding all the video files on the web is only part of the challenge. For video to be searchable, it is also necessary to collect meaningful text metadata to associate with each video file. Of course, we rely on standard techniques, such as mining closed-caption transcripts and importing RSS feeds. The vast majority of video on the web, however, does not have any closed-caption or RSS metadata available. Fortunately, our visual crawlers come to the rescue. Whenever our visual crawlers find a new video on the web, they can also “visually” examine the context of the surrounding web application. In most cases, this examination reveals a bounty of rich and detailed metadata related to every video.
With our unique approach to crawling, not only can we find the videos that the other crawlers miss, but we can also collect rich and relevant metadata for each video. As a result, when you search for video with Truveo, you can always find high-quality, relevant search results. The next time you are looking for a web video to watch, we invite you to try our search engine and see for yourself.
The only thing I’d like to see added is RSS feeds for search.
Dr. Timothy Tuttle – CEO & Founder
Dr. Adam Beguelin – CTO & Founder
Dr. Pete Kocks – Senior Architect