Alexa is opening up their index to whomever wants to mess with it. It seems they'll let you create and store programs on their servers to process the (100T! of) data, and query 'em via Amazon's web services platform...

The barrier to entry is a bit higher than that of Goog's or Y!'s API (both programatically, and 'cause there are modest fees), so it's prolly not going to encourage as many casual mashups, but the serious data mining it opens up is pretty staggering... What words contained in Joyce's Ulysses appear on more than 100 webpages? How many times does the word 'syncopated' appear on Portuguese pages with at least two .jpg's? How many vowels are there on the internets?