Bypassing custom CAPTCHA

[delirious]

Hi guys,


I'm running some project regards scrapping information about companies in my country. I have around 150 000 records in my database and it contains only a half of datas. Second part has to be scrapped from gov site.
Gov site is secured by custom made captcha: See the link below
https://prod.ceidg.gov.pl/CEIDG/CEIDG.Public.UI/Search.aspx


I wrote some script in Python in which i'm trying to OCR the image. But it does not work properly.
That's nothing illegal but it gonna be annoying to writing 150 000 captchas


Do you guys have any ideas ?


Does anyone can help me please ?

[delirious]

Yeah sure but i don't want to pay for that. I don't need any third party services. That is why i'm looking other ways

[delirious]

Can you tell me sth more about it ?


I wrote some script which was scrapping jpg with captcha then every digit was trimmed separately (except those which was sticked together) and compared with source.


But it was not working good enough.
Any help would be great.

[M1lak0]

I guess Using Regex would help bypassing CAPTCHA!

[Kulverstukas]

I guess Using Regex would help bypassing CAPTCHA!

Regex ain't shit if captcha is an image. However there was this one moment where some dudes could bypass Google's reCaptcha by analyzing audio it provides for those who couldn't see that captcha.

[delirious]

Thank you for all your replies.
Yesterday i found something strange and amazing in one time. It solving my problem but when I was thinking about that  ...i think it's impossible without internal access. Look at this video:


https://www.youtube.com/watch?feature=player_embedded&v=HLxruEi3h0M


This is exacly site i need to break captcha. What is strange ?
I think the author add some addon to web browser which ..."capture" request and response...between client and server.


You probably has more knowledge, experience than me. Can you tell me please is this possible ? and how it can be build ?

One fast question at the end. Fiddler is able to show all advanced "conversation" between server and client ??

Thanks for all your reply.

[techb]

A way to bypass captcha I seen on HaD was from the audio some give. If it gives you an option to "listen" to it, breaking it becomes much much simpler. No more image processing, just some audio filters and speech to text stuff.

Some captcha stuff HaD has posted, I'm sure the audio post is in there somewhere, but a resource non-the-less:

http://hackaday.com/?s=captcha