Hi guys!
I've noticed a lot of phrases like:
"we should give them a chance"
"this software has a great future" etc.
BUT how you can prove what you do not use OCR public modules?
I do not say about all captchas. Lets talk about ReCaptcha. Many people noticed your only because of ReCaptcha automatic recognition.
But in public exists ReCaptcha recognition module with
the same recognition params as you are offering.
The algorithm was introduced on the hacker conference long time ago. And you can find some existing realizations for example
http://www.antirecaptcha.com/
I think this is a very important questions for your clients, because if you can't write algorithm of ReCaptcha recognition by your self, you deceive your clients. If they know this, they would not expect so much from you and would not consider you outstanding programmers.
All in all your politics of the captcha recognition is very strange. You moved all captcha recognition codes to your server (not only ReCaptcha, but all others algorithms). These algos are all public and there is no sense to hide them.
You do not publish types of automatic recognized captchas. You do not publish any recognition rates. You do not publish any recognition speed. If it were yours algorithms, huge amount of them and their parameters you should not hide. But if there is no your code or just a little, it is logical to hide your code on the server to protect it against clients eyes or you even hire some guys to make human recognition.