Imagine scanning status updates for the topics you care about the same way you follow friends on Facebook, and you’ll get the idea behind Wavii, which crawls the web to deliver real-time personalized news feeds to its users.

There’s some pretty complex technology behind a service like this one, based around natural language processing (NLP) and machine learning.

That’s because compared to Facebook’s content, which is nicely structured (as status updates, likes, wall posts, etc.), content on the world wild web is largely unstructured, lacking the metadata that would allow a computer to recognize and contextualize it as accurately.

So Wavii is trying to solve that.

“We teach machines to learn exactly the way adults teach children to learn natural language,” says Wavii CEO Adrian Aoun. “It’s neuroscience applied to machine learning.”

When you sign in to Wavii at Facebook, it will prompt you to select an initial set of topics or people to follow – typically the bands, politicians, celebrities, brands or companies you’ve indicated you like at Facebook.

Then Wavii scours the web to deliver up-to-date content about these subjects to your news feed in a format very similar to that of Facebook itself.

“On Facebook, the cognitive overhead is nothing, so we're able to keep up with thousands of friends at a time, because Facebook is so good at context,” says Aoun. “I love this method of consuming information. But I'm frustrated that the web's information doesn't come in this format.”

In order to address that problem, he says, “We're making Facebook out of Google. We take the Facebook UI and apply it to the Google corpus. That way, we make staying up to date on the web so much easier.”

One of Wavii’s cool features is a “time machine.” Last week, as the Olympics were wrapping up, users could watch the gold medal race unfold between the U.S. and China hour by hour, and see how the U.S. literally ran away from the Chinese during the track and field events late in the week. 

Wavii’s technology gets smarter the more you use it, including when you indicate you like or dislike an item, share it or add comments.

It is also available as an iOS app.

The company, which launched this past April, is based in Seattle, but has close ties to the Bay Area. “We are a Bay Area company that happens to be based in Seattle,” is the way Aoun, who visits the Valley pretty much every other week, puts it.

The company’s investors include Ron Conway/SV Angel, Aydin Senkut/Felicis VC, Mitch Kapor/ Kapor Capital, Mike Arrington/CrunchFund, Dave Morin, Shawn Fanning, Keith Rabois, Joshua Schachter, Paul Buchheit, Max Levchin, and Rick Marini, CEO of BranchOut, the professional network tied closely to Facebook’s platform.

Most of Wavii’s 35 employees also have Bay Area ties, including Stanford grads, ex-Zynga and SoundFlavor employees. In addition, over the next year, the company is expected to open an office in San Francisco.

The demographics of Wavii’s early users break down roughly 60-40 male/female, and 50-50 U.S. and overseas. Aoun says the average session per user at launch was six minutes, and now it has grown to eleven minutes.

It’s still early, of course, but that sounds like Facebook–type stickiness to me.

Note: Please check out of new ebook, 30 Startups to Know Now.