I tried it by providing url’s for several different profiles and noticed that the results varied but were still impressive. I’m guessing that that the discovery process requires the same username across services which is the case for me. Entering my Twitter profile url discovered 21 profiles and FriendFeed discovered 29 profiles. It will also pull a different profile descritptions for you depending on the site you used.
Example of my results from the demo page. Click image to try it out.
There are several other app demos available that pull profile info along with data as well as one that pulls profile photos across services. Beyond support for the 70 sites the library also supports a multitude of content definitions as well as data schemas. The engine is already available for download and carries an MIT license. I am extremely impressed by this script and could see people using it for self-hosted lifestreams very soon. I wouldn’t be surprised if someone created a WordPress plugin utilizing this library soon as well.
You can read more about Ident Engine over at ReadWriteWeb as well.