Their OCR did worse at reading a 40 page PDF of printed text than tesseract for me, I just paid $2 for useless output