The other day Jackie and I were wrapping up a podcast and he mentioned something about using “stream” instead of a file. I mentioned that I’d adapted my OCR script to use the built-in Windows 10 OCR (instead of relying on Tesseract) but I was unable to change my script to use a variable instead of requiring a saved file. Thankfully Jackie had actually played with this so we started to update my script. About 30 seconds in, I suggested we record the session and I’m glad I did as there was some advanced stuff we covered! We ended up turning it into our 55th Podcast.
In this webinar on getting text from programs Jackie Sztuk demonstrated his adaptation of Tesseract. I realized this would be a cool addition to my screen-clipping tool so I’ve incorporated! Check out the below video and you can download all the files for screen clipping with OCR here (This requires several executables so I’ve put it all in this zip file) BTW on 9/21/19 I updated the Tesseract.ahk file to wait for the image to be present. This should reduce the issue of the script complaining about the image not being found.