I wouldn't hold my breath on it. Real-world OCR is hard. According to my tests, even Microsoft, Google or IBM's APIs get it wrong very, very often.

not that I've seen, though with some logical combination of object dimensions, env variables (depth of field, etc), and a decent image recognition/ML then you could probably get decent OCR. Note that this is for visual recognition, feel text is decently solved (exhibitA, Google Translate).