Creating a workflow to recognise text in an image

[nikivi]

I had an idea of making this workflow for some time now. I want to have a screenshot utility that will take a screenshot created by a user, it will recognise the text in this image and put the text into the clipboard of the user. 

 

So if I was to screenshot this line from wikipedia for example : 

 

 

 

It would scan it, and put a text representation of this image into my clipboard. So in my clipboard after the scan, I will have this : 

 

 

 

I found a library that seems to do what I want : https://github.com/otiai10/gosseracts 

 

However I am not quite sure how to connect all these pieces. I can make a screenshot by issuing a screencapture -ci command. Then perhaps I need to save this image to some temporary location and then run this or some other OCR program on it and get the text I need?

 

This workflow is specifically useful for me for grabbing text from places where text is non selectable. Like for example, I wish I could screenshot this image : 

 

 

 

 

And get the text from this image into my clipboard.

 

Thank you for any help.

 

 

Edited January 28, 2017 by nikivi

[blizzrdof77]

You could probably use google docs API? Or this android vision API:
https://codelabs.developers.google.com/codelabs/mobile-vision-ocr/#0

[Bhishan]

You can use Mac app   Mathpix snipping tool .

 

It gives Latex version of OCR,  but you have to convert latex to plain text manually.