[rjzzleep]

OCRmyPDF does that. Given the name I assume it's only for PDFs given the name, but theoretically it should work for any image without the need to extract it from the PDF.

https://github.com/jbarlow83/OCRmyPDF