takeheart , vor einem Monat

I mean taking the screenshot is the easy part, getting reliable OCR on the other hand ...

In my experience (tesseract) current OCR works well for continuous text blocks but it has a hard time with tables, illustrations, graphs, gui widgets, etc.