AdaptiveBlue’s latest version of its browser plug-in is a complete revamp of its BlueOrganizer product. Now called Glue, it aims to synthesize semantic technology that connects information about books, music, movies and similar subjects with the browsing habits and commentary of your friends.

Take a book listing on Amazon.com, for example. AdaptiveBlue’s plug-in directly injects links and information into the web pages you’re browsing without having to move to another site. So you’ll be able to see the parallel listing for Barnes & Noble, and any Wikipedia entry on the book, to name a few. What defines AdaptiveBlue’s new product, though, is the ability to add friends and other surfers to the mix.

With Glue, if any other users have visited that same book, whether on Amazon or another site, the plug-in will flash a bar showing them. You can then see where else others have been browsing — completely, anonymously of course. Users can also say whether they liked the book, and leave comments if they want. Though the assumption is that most users will remain passive most of the time.

Glue will only track the last 20 objects any user visited, but it could build up information over time on what users like and dislike, as well as provide a sort of recommendation network. The one problem for AdaptiveBlue is that — because most of what it follows are products (stocks are a new exception) — the companies selling them already include their own recommendation and comment sections.

In the Amazon example, the company not only suggests similar books, but also offers user-created lists. These suggestions are a vital part of the marketing strategy for companies like Amazon to become more relevant and useful.

All Glue offers in addition is the chance to see what your friends are interested in, provided they also use Glue, and only if you’ve become friends with them on the service. Given all these caveats, and the low added value to start (although Glue could build up useful info over time), it doesn’t seem likely that people will flock to the social aspect of Glue.

Still, founder Alex Iskold says that the 600 beta testers the company used were “very excited,” and that there are plenty of sites that don’t offer recommendations where Glue might come in handy. Ultimately, whether people will want to use Glue or not will come down to surfing habits, and the plug-in does have a smooth, easy-to-use interface going for it.

To date, about 1.7 million people have downloaded AdaptiveBlue’s BlueOrganizer plug-in, and 350,000 of them are active users. They’ll probably be migrated to Glue within about a month. The New York company also recently raised $4.5 million, led by RRE Ventures and with participation from Union Square Ventures.

Semantinet, a company that hopes to give users more information about topics on web pages without them having to perform a search or leave the site they’re on, is launching a product called Headup into private beta this morning at the Web 3.0 conference in Santa Clara, Calif.

People who have been following similar startups will see an immediate comparison to another company, called AdaptiveBlue, because Headup also requires the download of a plugin for Firefox. That gives it the ability to directly modify web pages that you’re viewing in the browser (a few other companies, like Zemanta, also use this approach).

The idea is to give on-the-spot access to reams of information that today’s web sites can’t or won’t add for themselves. So if you see a reference to a movie on a blog, for instance, a tiny Headup icon will appear beside it. When you click the icon, a pop-up box will appear that might show, for example, clips from the film on YouTube, a review posted on another blog, or a comment a buddy left about it on Friendfeed.

Startups like Semantinet are in effect trying to take a shortcut to the next generation of the Internet, by forcing additional content into web pages — what’s often called a “top-down” approach to the semantic web, rather than the bottom-up approach of the web site owners structuring their own content so that machines can easily read it and link it to other sites.

The problem that top-down startups avoid is the painstakingly slow development of the semantic web; most data is still just plain text that can’t be easily accessed by other websites. However, a top-down approach has its own dangers. The biggest is that users are forced to download a plugin, which immediately excludes the majority of the Internet.

Those users who are motivated enough to download Headup will require a careful balance between being deluged with useless information and having a plugin that doesn’t do enough. AdaptiveBlue’s BlueOrganizer, for example, is good at giving information about things like books and music. But for a web surfer who only occasionally reads about those things, it may not seem worth keeping installed.

Headup is taking aim at similar subjects, although CEO Tal Keinan says that others, like people and blog posts, are also part of the package; on a blog, for example, Headup might be able to give you a preview of a story before you click on the headline, or it might show information from Wikipedia when you ask it to expand on Warren Buffet’s name. But the real target audience for the moment are tech-obsessed geeks, with a number of friends online.

That’s why Headup has already been optimized for sites like Friendfeed, and given the ability to aggregate information about friends from networks like Facebook, MySpace and Twitter. Click beside a friend’s name on any of those, and Headup will show you what they’ve done on the other sites. Of course, Friendfeed itself tries to do something similar, so the question becomes which interface nets in all that information more attractively.

Over time Semantinet aims to become attractive to people beyond the early adopter crowd, by working work progressively to bring in ever more subject areas. One example could be company names; Headup could easily feed in more information about startups on a blog like VentureBeat, or stocks on CNN. But each of those category types has to be optimized to be truly useful, so expansion will take time.

This is the first time that Semantinet has talked openly about what it’s building. A month ago, the Israeli company also announced a $3.4 million second round of funding. Its first round of $1.7 million closed early this year.

The semantic web is on its way — or, according to Adaptive Blue CEO Alex Iskold, already here, and just waiting for the right applications to put the pieces together.

With the latest release of BlueOrganizer, a FireFox plug-in that provides an alternate way of browsing the web, he’s got a more compelling argument than ever.

BlueOrganizer is approaching the semantic web by consumer categories. Since its founding in 2006, it has built out data on subjects like blogs, books, movies and restaurants.

If you had the plug-in and I mentioned, for instance, the book “A Thousand Blazing Suns” by Khaleid Hosseini, BlueOrganizer would highlight the name, provide further information and links to buy the book, learn more about it or share it with others.

New to the app is the ability to access micro=formated info like a name and address. Right click anywhere within it and you’ll be able to automatically map the location or check out nearby restaurants.

Or, for a person, right-click on their name and be able to shoot off to their Facebook or Twitter profile, look them up on Wikipedia if they’re famous, and so forth.

Iskold calls it a “top-down” approach to the semantic web. Rather than tagging data before it’s placed on the Internet, BlueOrganizer looks at what’s already present and uses it to help people in their ordinary browsing activities.

In the process, it’s re-structuring the way its users navigate the web by providing a browser-based overlay to websites that users navigate by, rather than the links alone within pages.

The company took $1.5 million from Union Square Ventures in early 2006, and is currently considering another round.