simpsons corpus

i heard a story about an MIT project where an eliza-styled program answered queries with quotes solely from star trek or the simpsons.

They must have built or found a massive corpus of simpsons quotes.
Does anyone have any idea how this was done? I imagine it’s a few scales of magnitude larger than the odd “simpsons favorite quotes” files on the web.

Thanks!

This would probably be a good place to start.