I can't believe there is a common interface. I can try to explain why. Can you keep a secret? :-) After looking through several OCR source code, I got the impression that many teams can be very good is some specific fields of Applied Mathematics while having no idea about programming. Nevertheless, they do programming and do what they do. So, some teams have no enough culture to understand interfaces and other stuff. In such cases, the code is very much proprietary, integration with other software is ugly or intrusive, etc. For example, I was really disappointed with support of different languages and Unicode, it looks like authors just did not know how Unicode works. Again, my experience is very limited and based only on Open Source codes. I don't want to finger-point to any product at this time.
(By the way, I found some very mature works on CodeProject, but all I found was not complete to make a working application. Please find the reference in my past answers:
OCR Software[
^],
Apply online handwriting on neural network[
^], look at 7 CodeProject article references here.)
—SA