Is anyone out there willing to take on a real challenge? Then how about adding full text indexing into your aggregators! The parsing algorithms are well documented and simpler then you might think. The actual chokepoint is usually the database, which needs to be able to update and maintain fairly large word indexes... I got a C# prototype up and running under Access and/or MSDE if anyone is interested.