I thought I had found exactly what I was looking for: https://github.com/tesseract-ocr/tesseract
Everything was perfect, until I read:
Since 2006 it is developed by Google.
I would rather not have Google spying on everything I OCR, please.
Is there an alternative which is not infested by Google?
I've looked long and hard and it sadly doesn't seem to be the case. Although I wonder how exactly it's "developed by Google" since nothing about that repo seems to suggest that it is except for that sentence...