Parsing Photo ID engine

Hi, I have a website where a user will go to a form and enter his first name , last name, and then an email address, at one point we ask for a photo id ( a passport photo biometrics page). I need an engine that can be deployed on AWS, that will take the image, the first name, the last name as text, and do some image parsing recognition, in the following way : - The system should accept all photos, if the passport is from china or amerika, it should return an invalid message. - any other country, it should parse the english language section ( ie : saudi arabian passport has arabic and english) and makes sure the information the user entered about his first and last name match the passport image photo ( ie : Ali Ali is spelled exactly Ali ali in the passport photo), if that is correct it returns a message saying its ok, and then saves the image in an S3 Bucket ( or some DB on AWS). - if the names don;t match but its a misspelling it should return the parsed and the wesbite will suggest changing the name. - This should only work for english language, so any passport should work as long as there are english language available on it. I have looked at Python and tesserect but doesn;t look very promising, I am interested in this being an engine and no external API call. Please make sure you read the project description fully and respond with the word " OCR" in your bid so I know you read it fully.