I've worked with over 20 different OCRs over the last 15 years, most of them commerical OEM bits. Tesseract is an excellent choice, particularly being free. It is very fast and very accurate on good quality images. Scanned documents and smaller fonts tend to cause Tess to produce very poor results, but your typical font sizes and resolution above 72dpi Tesseract is my first go-to. The only OCR SDKs I've found better you paid for. You paid a lot. The problem I last left with Tesseract was the ability to extract multiple choices for glyphs. On both Ubuntu and Windows, 64bit and 32bit, the software has some serious crash problems using that API. I think that had I continued pinging the forum that problem would have eventually been resolved. The forum can be helpful if you have a few days to wait for the answer. You want a Free OCR SDK?? Tesseract is the answer.