Dominik Wujastyk's blog
That is fascinating. As far as I can see, though, the concordances and analysis are only available for each individual text. That's a pity, because their features would be insanely useful if they were applied to an entire corpus, or if you could easily compare texts to each other or to different sections of a corpus. That would open up so many possibilities in terms of dating texts, tracking changes in word meanings between language communities, identifying later interpolations, guessing at authorship, etc, etc, etc.
It's true. Then again, they do let you suck the indexed texts down to your local workstation. So it would be possible to develop local tools for more advanced analysis.