EveryZing, a search startup using speech-to-text to analyze what’s inside video, is launching a product to help media sites promote their video content. The company had switched to a business-to-business focus when it raised a $10 million Series B round last June, and new product “ezSEO” is the fruit of those labors. Its original podcast search engine is now a “a good proof of concept and sandbox but not any part of our revenue success,” according to EveryZing CEO Tom Wilde.

The problem, as usual, is it’s hard to find out what’s going on in a video so it can be searched and advertised against like text. EveryZing’s solution is to make video act more like text. For customers such as Boston.com, the company creates transcripts of videos and then maintains search engine optimization (SEO)-friendly pages of text related to them. Google has added 50,000 new Boston.com pages to its index in the last month, driving a fresh heap of traffic to the site’s video. I don’t see why any media company with a video archive shouldn’t give this a try.

EveryZing, which also provides customers with a video search engine, charges a monthly service fee and a performance fee for search page result views. The products are mostly automated, but require maintenance of a custom dictionary for speech recognition, so, for instance, the search engine doesn’t get flummoxed by a sportscaster’s mention of a new Red Sox player’s name. EveryZing ingests video from RSS feeds so as to make integration as easy as possible.

The logic is simple, said Wilde. “If users can find more, they’ll consume more. If users consume more, there’s more advertising opportunities.”

EveryZing spun out of research firm BBN Technologies, and was previously called PodZinger.