> Tesseract is an open source text recognition (OCR) Engine, available under the Apache 2.0 license.. OP here. In my testing it's only reliable for images with relatively clear text at a 90 degree angle. Results may be improved by cropping non-text areas of the image. I didn't try any extra data sets or non-English language packs.