[PHP][Source Code]antirecaptcha

Discussion in 'Programming' started by p3n1, Oct 10, 2010.

Thread Status:
Not open for further replies.
  1. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    I've written a recaptcha distortion remover and now i'm able to download with jDownloader from HF and FS (and most likely from all other sites using reCAPTCHA).

    New Version 0.2c - got 17% recognition rate!

    the combined module for jDownloader can be downloaded here.
    hope u can use this stuff ;)

    # anti-recaptcha v0.2c (c)opyleft 2010 http://wegeneredv.de/arc **************
    # ****************************************************************************
    # future fixes and changes:
    # - gaining greater recognition rates by using wikipedia as wordlist
    # - see included paper 'antirecaptcha.pdf' for further information or
    # download it from http://wegeneredv.de/antirecaptcha.pdf
    # - clean up source / replace repeating code by functions
    # - convert to c++ using http://github.com/facebook/hiphop-php
    #
    # ****************************************************************************
    # Installation:
    # copy directly to your jDownloader installation directory (e.g.
    # c:\program files\jDowloader and extract.
    # -runs only on Windows systems-
    #
    # ****************************************************************************
    # Changelog:
    #
    # v0.2c
    # - applied selective contrast to both words
    # - improved success rate to 17% by adding a bigger wordlist to tesseract
    # - some hoster aren't supported any longer due to plugin errors caused by jD
    #
    # v0.2b
    # - removed 'best match' for speed reasons
    # - removed intersection cleaning due to ineffectivity
    # - added experimental sharpening
    # - changed from VietOCR to tesseract
    #
    # v0.2a
    # - pushed recognition rate to ca. 10%
    # - multiple iterations to ocr ('best match')
    # - partially cleaned intersections between letters
    # - cleaned up sourcecode
    # - added support for the following onc-click-hoster:
    # combozip.com,cramit.in,crazyupload.com,drop.io,enterupload.com,extabit.com,
    # filebling.com,filechip.com,filescloud.com,fileserve.com,filesmonster.com,
    # filesonic.com,filestab.com,filestrack.com,filevo.com,freakshare.net,
    # free-share.ru,hatlimit.com,hidemyass.com,hostingcup.com,maknyos.com,
    # mediafire.com,putshare.com,quickupload.com,slingfile.com,tgf-services.com
    # uploadfloor.com
    #
    # v0.1b
    # - changed the directory structure
    #
    # v0.1
    # - first version tested
    #
    # known bugs:
    # - hotfile.com folder gets deleted by jDownloader:
    # -> set NTFS-permissions to 'read only' for the user group 'everyone'
    # - sometimes the captcha input window pops up - just *ignore* it
    cah and Raspbooty like this.
  2. xhane

    xhane New Member senior old school member

    Joined:
    Jun 8, 2009
    Likes Received:
    35
    Location:
    Up yours
    Thank you for this. Not trying to flame, but since you've posted it, Google will probably fix it very soon.
  3. cah

    cah New Member old school member

    Joined:
    Jun 11, 2010
    Likes Received:
    7
    Thanks. What fraction of captchas can it decode?
  4. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    if you mean by fraction which recognization rate it has: around 5%... thats not that much but i'm working on it. and should be enough if you use proxies because of the ~32 wrong guesses google gives you.
    if they fix it, i got something to do - and lets be clear: i totally want them to get better, cause its the only spam stopping mechanism out there atm.
  5. dest

    dest New Member coldschool senior old school member

    Joined:
    Sep 26, 2009
    Likes Received:
    98
    [​IMG]
    ↓↓↓ ↓↓↓ ↓↓↓ ↓↓↓ ↓↓↓ ↓↓↓ ↓↓↓ ↓↓↓

    [​IMG]

    it seems to me that the result is far from being impressive, not bad though

    Attached Files:

  6. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    thx for your reply! you're absolutely right, it could be far better - that's what i hope for. i haven't implemted all of chad's algorithms yet and the parameters
    Code:
    $contrast=128;
    $maxslope=-1;
    $mid_exponential_smoothing_factor=0.2;
    $vector_exponential_smoothing_factor=0.001;
    
    aren't the best fit by now. but i think it's a good point to start from.
    but feel free to play around with them and post your results afterwards!

    little update:
    i get better results (around 9%) with these parameters:
    Code:
    $contrast=112;
    $mid_exponential_smoothing_factor=0.1;
    $vector_exponential_smoothing_factor=0.04;
    
  7. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    Update to v0.2a available:

    Changelog:
    v0.2a
    - recognitization rate pushed to 10%
    - multiple iterations to OCR (“best match”)
    - intersections between letters partially removed
    - cleaned up sourcecode
    - added support for these oneclick hoster:
    combozip.com
    cramit.in
    crazyupload.com
    drop.io
    enterupload.com
    extabit.com
    filebling.com
    filechip.com
    filescloud.com
    fileserve.com
    filesmonster.com
    filesonic.com
    filestab.com
    filestrack.com
    filevo.com
    freakshare.net
    free-share.ru
    hatlimit.com
    hidemyass.com
    hostingcup.com
    maknyos.com
    mediafire.com
    putshare.com
    quickupload.com
    slingfile.com
    tgf-services.com
    uploadfloor.com

    more info: http://bitland.net/captcha.pdf and http://n3on.org/projects/reCAPTCHA/
  8. rpz79

    rpz79 New Member

    Joined:
    Oct 17, 2010
    Likes Received:
    0
    GJ Man

    At last someone who reallly tries to deal with this annoying problem.

    I hope you'll soon reach at least 50% hit rate :rolleyes:

    Currently in v0.2a hotfile.com ACR isn't working for me he's asking to enter it manually.

    KEEP UP THE GOOD WORK BUDDY
    :)
  9. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    thx for the motivation :)
    but i think 50% will never be reached - there are too many variations...

    jDownloader deletes the 'hotfile.com' folder everytime it starts, so you'll have to set the ntfs permissions to read-only for the usergroup "everyone" or just take another hoster, because hotfile got a 1 min delay between captchas ;)
  10. thepenguin99

    thepenguin99 New Member member

    Joined:
    Sep 26, 2010
    Likes Received:
    0
    Pretty useful if you have the proxies. That said decaptcha is pretty cheap so not sure I will make any use of it.
  11. rpz79

    rpz79 New Member

    Joined:
    Oct 17, 2010
    Likes Received:
    0
    Could you PLZ explain , what proxies are you talking about ?

    DeCaptcha is the site decaptcher.com ? he deals with google's ReCaptcha aswell ?

    THNX

    Yes SIR !!! XD
    First Time Hit both for FS and HF !!! XD
    FANTASTIC JOB :biggrin: Just try Keeping those %s UP my dear friend :biggrin:

    Listen for the meanwhile isn't there an option to return to manual (user)
    AntiCaptcha after X auto AntiCaptcha mishits ?

    THNX ALOT
    ;)
  12. www

    www New Member

    Joined:
    Oct 18, 2010
    Likes Received:
    0
    pretty fuckin nice
  13. sublimesed

    sublimesed New Member

    Joined:
    Sep 13, 2010
    Likes Received:
    0
    sorry for being a noob but can anyone help me set this up in English?
  14. coldasice

    coldasice New Member

    Joined:
    Oct 19, 2010
    Likes Received:
    0
    thanks dude this helps em play a game and dl :))))

    keep the % and recive internet love
  15. rpz79

    rpz79 New Member

    Joined:
    Oct 17, 2010
    Likes Received:
    0
    To add the module to jDownloader, the archive must be extracted to the folder "installation folder * * jdownloader \ jd \ captcha \ methods \

    ;)

    for the meanwhile return to manual (user) AntiCaptcha after X auto AntiCaptcha mishits ?

    Currently I have to cut out HF and FS folder out to do so.

    We need the %s up at least to 25%.
    FS (and maybe HF) gives a plugin error after +/- 10 failures and move on to the next file so there goes my night DL :eek:h:

    I don't know how you work on this but if the algorithm isn't common to all hosters I really really suggest that you focus on the main 2:
    FileServe and Hotfile. Anyway how are things progressing ?

    I'm really crossing my fingers for you :blink1:
  16. ggcollins

    ggcollins New Member

    Joined:
    Oct 19, 2010
    Likes Received:
    0
    Sorry I'm a noob here but can you explain me step by step what should I do to use it for a specific website.

    Thanks ;)
  17. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    this should help

    ---

    i wrote a little paper about the process of antirecaptcha... :noworry:
  18. namhoang

    namhoang New Member senior member

    Joined:
    Mar 18, 2010
    Likes Received:
    39
    any1 can make it work with jdownloader? I tried but it seems doesn't solve the captcha for filesonic
  19. rockmore

    rockmore New Member member

    Joined:
    Sep 4, 2010
    Likes Received:
    66
    yeah it didnt work so well for me
  20. p3n1

    p3n1 New Member

    Joined:
    Oct 8, 2010
    Likes Received:
    2
    filesonic doesn't work with jdownloader atm - their plungin is outdated, it got nothing to do with antirecaptcha ;)

    new version is out btw :biggrin:
Thread Status:
Not open for further replies.