If you point Universal Summarizer at a PDF which is scanned text it will fail with the "input too short" error.
While technically correct the error is misleading. There is plenty of content, you just don't OCR it.
As a first step a better error message would be something along the lines of "we can't find any content to summarizer" but the better answer would be to OCR it.
I'm using a deskstop package which did a pretty good job and then fed that to the summarizer, so it is possible.
