Don't think you can. And also there is big difference in plain old OCR, which is just getting all text out from image and document processing which is can you only get the relevant information in a good structure that can be directly pushed into a database.
