Right, I forgot that they also read and parse text in images well.