The author uses the terms API and "indexing technology" imprecisely here. I think what he really means is that they are getting data from Twitter via a <i>push</i> protocol rather than a <i>polling</i> protocol.<p>Note that while Twitter has a site-specific protocol (the <i>firehouse</i>), any site that wants to enable real time searching and alerts can also just implement PubSubHubbub, which can notify any interested subscriber about updates in near real time.<p>As far as I know, Google uses PubSubHubbub to update Google Reader feeds in real time - I'm not sure if this integrates into realtime search results as of now.