Python voice comparing, words recognizer

I am looking for some open source tool to solve my problem.

I need to compare two or more sound files spoken by the same person. I am not talking about voice recognition which I am using in another part of project (wit.ai and cmusphinx). What I need is to record a child (which did not speak correctly due to some mental disabilities) speaking fe. Single word, and then ask him/her to speak the same word again. And if he/her would not pronounced it correctly (with some precision of course) an error would be raised.

I would be gratefull for any help :)