This was churning in the back of my head, so I toyed with it a bit. wget http://doublerecall.com/images/mockups/DoubleRecall-iPod-engagement-explainer.png

This was churning in the back of my head, so I toyed with it a bit. wget http://doublerecall.com/images/mockups/DoubleRecall-iPod-engagement-explainer.png gocr -i ./DoubleRecall-iPod-engagement-explainer.png -m 4 -C A-z -d 1 Nets me: __ ___ __ _______ ____ ___ __________ ______ ___ e_ n _ Apple iPad ___l_J__ Movies_ TV shows_ games_ and music_ Now pIaying on an iPod near you_ Now pIaying iPod I _ ce_ con_en_ \ Throw out the non-recognized characters, and I've got a list I can grep against a dictionary and I've detected all the required words already. It wouldn't take much glue to call up imagemagick and detect color differences between word chunks and fill this in automatically. Chrome Extensions can capture page content to canvas datauri's and upload to a server for analysis.