From: University of California at Santa Cruz
MetaVid (http://metavid@ucsc.edu) is an open source project from the University of California at Santa Cruz that allows users to keyword search EVERY word spoken (aka transcript search) on the floors of the U.S. House and U.S. Senate and then listen to the audio and/or watch the video. The text transcript comes from the closed captioning associated with each file, and the index dates back to January, 2006. Files can be viewed online and downloaded if desired. Search terms can be set-up on an RSS feed, and an advanced search interface is also available