[Solved] Do captcha's get sent using the proxy settings

apparently for some reason google hates my IP. So I bought 4 socks proxies to download with inside Jdownloader. With 4 different IP's I cannot get one right answer on filejoker. Are the captcha's submitted using the socks IP entered into jdownloader?

[pspzockerscene]

Well if jd has to open your browser for the captchas I guess the proxies won't be used for that (but I might be wrong).

If you added your Google account to JD, Google might still hate you BECAUSE you're using proxies and they recognize this (though I'm not sure about this either).

Google will at least notice that you use one Google account with multiple IPs at the same time.

If now, other users use the same proxies (same IPs) with their Google accounts (chances are high) it might look like BOTs for Google --> They will give you hard/unsolvable captchas.

GreeZ psp

Okay that sounds right. Because I got a working captcha before I added my google account to Jdownloader. I will try to remove it.

Also is there any way that the in browser captcha can be sent with the proxy IP or sent to captcha service.

Because currently if your user-agent + IP is blocked by filejoker you are stuck. No amount of correct captchas will start a download.


Only way to fix this is to change user-agent with a tproxy. Other plugins allow user -agent switching.

I cant change my ip so I bought private proxies and it still is not working. Allow us to change the User-agent or else this is never going to work. Buying proxies does nothing if your home IP is still being used for captcha.

I have root on these VPS's no one else can use this proxy. It's simply not working because Jdownloader is leaking the real IP still.

[raztoki]

Filejoker if its still using Recaptcha,
and your settings are to use browser loop to gain captcha quality advantage, then when JD fires a browser event, its totally in the hands of your browser to open this session. JD fires off open URL function to your OS API or to the path binary with switches you provide, if your browser allows to say open URL with a given connection we might be able to support that. I must say I use proxy a lot in JD and not in browser, and never had issue. Far as I understand hoster doesn't know the IP address that has been communicated to Google API for Recaptcha.

Where as if Google account is provided, you could have issues with login if accessed a lot from different devices simultaneously, though I would add here that it shouldn't be impossible to use. Since these days one can have series WAN connections even at home (mobile[phone|tablet] && physical), users also be using VPNs or VPS which would mean concurrent connections from multiple locations. Google does track logins from user-agent and API addresses and notifies Google account of new login from given device to ensure account holder that hacking or login from interface that wasn't them.

Originally Posted by raztoki

Filejoker if its still using Recaptcha,
and your settings are to use browser loop to gain captcha quality advantage, then when JD fires a browser event, its totally in the hands of your browser to open this session. JD fires off open URL function to your OS API or to the path binary with switches you provide, if your browser allows to say open URL with a given connection we might be able to support that. I must say I use proxy a lot in JD and not in browser, and never had issue. Far as I understand hoster doesn't know the IP address that has been communicated to Google API for Recaptcha.

Where as if Google account is provided, you could have issues with login if accessed a lot from different devices simultaneously, though I would add here that it shouldn't be impossible to use. Since these days one can have series WAN connections even at home (mobile[phone|tablet] && physical), users also be using VPNs or VPS which would mean concurrent connections from multiple locations. Google does track logins from user-agent and API addresses and notifies Google account of new login from given device to ensure account holder that hacking or login from interface that wasn't them.

Yeah I only have this problem with filejoker. And it started after my imagetyperz account got a few google captchas wrong. Now my IP+jdownloader user agent is blocked. No amount of manual captcha solving will start a download inside of Jdownloader. Browser will still work so my IP is not completely banned.

Here is what I just did to fix my issue. Maybe something can be worked out or filejoker plugin given the ability to change user-agent like vk plugin.

I logged into all 4 of my vps's. Installed a HTTP Proxy as transparent. Set it to modify the user agent to chrome on every header.

then deleted my google account inside of Jdownloader. Then pointed all the proxies inside of jdownloader to use the http proxy with the modified headers and different user-agent string. Proxy inside of browser also pointed to modified HTTP proxy.

Now all 4 downloads started perfect. I also got the easy captchas again.


So it seems filejoker has access to google's ip's of the submitted captchas and stores a browser fingerprint of IP+user-agent.


Would be nice if Jdownloader sent every captcha for each download to google with the correct proxy ip instead of using the real machine IP.

[raztoki]

far as I know google doesn't give that sort of feedback to clients,
api key in hoster source > client opens and communicates with google > user solves > submits response to hoster > hoster confirms challenge id & response with google > google send ya or ney.

JDownloader uses the proxy selected within candidate, the browser from the plugin is set to the captcha class. If you rather not use real IP take it out of use (uncheck no proxy).

raztoki

Originally Posted by raztoki

far as I know google doesn't give that sort of feedback to clients,
api key in hoster source > client opens and communicates with google > user solves > submits response to hoster > hoster confirms challenge id & response with google > google send ya or ney.

JDownloader uses the proxy selected within candidate, the browser from the plugin is set to the captcha class. If you rather not use real IP take it out of use (uncheck no proxy).

raztoki

I had no proxy unchecked. If google is blocking you with IP+user-agent filejoker will never give you a correct captcha response. If right captcha is entered they still will not allow you to download. Someone else of this board PM with that exact same problem.

Only way to fix is change IP or user-agent.

I cant change my home IP so tried to use proxies but that still didnt work. Only fix is to rewrite header of user-agent which is a lot of work.

VK plugin allow user agent switch I dont know why Jdownloader dont want or allow this.

[raztoki]

user agent is already random, thus no point in making any user preference. All you need todo is restart the client when blocked.

raztoki

Originally Posted by raztoki

user agent is already random, thus no point in making any user preference. All you need todo is restart the client when blocked.

raztoki

I was told different here https://board.jdownloader.org/showthread.php?t=66583

Also spent days on this and redownloaded Jdownloader, nothing fixed the captcha being blocked until IP or User agent was changed.

During all this filejoker worked inside of browser but would never download in JDownloader. another user PM me with the exact same problem.

[raztoki]

I checked the source code when I made my comment, just double checked just now

    private final boolean                  ENABLE_RANDOM_UA             = true;

checked history, it was made random with revision 27797 2014-11-05 03:58:37 AWST by psp. Nearly a year ago!

user-agent plays very little influence with recaptchav1 anyway. It's mainly cookie based. This is why using browser loop using your cookies from a standard browser or google.com account assists. Since you every day search engine wise or what ever, you build up reputation that your not a computer... You can easily destroy that reputation by requesting too many captchas in short time or answering the wrongly, thus google thinks your a bot... Browser loop is better in my opinion as you can at least flush you cookies and start again, but with account you can do (maybe) irreversible damage.

raztoki

Originally Posted by raztoki

I checked the source code when I made my comment, just double checked just now

    private final boolean                  ENABLE_RANDOM_UA             = true;

checked history, it was made random with revision 27797 2014-11-05 03:58:37 AWST by psp. Nearly a year ago!

user-agent plays very little influence with recaptchav1 anyway. It's mainly cookie based. This is why using browser loop using your cookies from a standard browser or google.com account assists. Since you every day search engine wise or what ever, you build up reputation that your not a computer... You can easily destroy that reputation by requesting too many captchas in short time or answering the wrongly, thus google thinks your a bot... Browser loop is better in my opinion as you can at least flush you cookies and start again, but with account you can do (maybe) irreversible damage.

raztoki

Why would recaptcha work inside of the browser but deliver 100 percent wrong captchas even if their right inside Jdownloader? My IP never changed and I tried with and without a google account inside of Jdownloader.

According to here https://board.jdownloader.org/showthread.php?t=66583

As soon as I changed the user-agent to chrome recaptcha again let me download from filejoker.

This only happened on filejoker. Other hosters that used google captcha worked.


Another user has the same problem and we both are using Jdownloader on linux.

[raztoki]

Few people have issues, either with miss configured system (browser loop not working, thus reverts to old recaptcha handling, gets shit image quality and then answers are nearly always incorrect/impossible to solve.) or even code issue on our side (I did Teamviewer session and found NPE caused problem), seen other issues.. its hard to say without reading more into your logs which are not been influenced by proxies since the headers are been manipulated (thus not true).

once again that thread is full of misconceptions or rubbish, since I proved above that random ua been used already for nearly a year. It should play little to no influence, you can achieve the same outcome by restarting client.
raztoki

Originally Posted by raztoki

Few people have issues, either with miss configured system (browser loop not working, thus reverts to old recaptcha handling, gets shit image quality and then answers are nearly always incorrect/impossible to solve.) or even code issue on our side (I did Teamviewer session and found NPE caused problem), seen other issues.. its hard to say without reading more into your logs which are not been influenced by proxies since the headers are been manipulated (thus not true).

once again that thread is full of misconceptions or rubbish, since I proved above that random ua been used already for nearly a year. It should play little to no influence, you can achieve the same outcome by restarting client.
raztoki

which log are you referring to? Are you saying that the headers are not matter? Because I have reinstalled Jdownloader since that first log was submitted.

raztoki i tested it without the header modified. No proxy disabled. recaptcha gave me 3 captchas. All submitted manually by me. I made 100 percent sure they were correctly typed. After the third captcha the download started.

So probably my fault somewhere thank you.

[raztoki]

@cheeseand
logs are within jd2/logs/session/

one I would be interested in would be filejoker related

you can generally tell which is browserloop/account vs non, you get house numbers vs burly quality images which are next to impossible to solve.

raztoki

raztoki

Originally Posted by raztoki

@cheeseand
logs are within jd2/logs/session/

one I would be interested in would be filejoker related

you can generally tell which is browserloop/account vs non, you get house numbers vs burly quality images which are next to impossible to solve.

raztoki

raztoki

Im in linux so I think the browser loop wasnt working for whatever reason. Thus my Imagetyperz account got a bunch of captchas wrong.

1. I redownloaded Jdownloader
2. Install chromium
3. Set chromium as default browser in linux
4. logged into google account inside of chromium
5. deleted google account inside of jdownloader
6. enabled all my proxies inside of jdownloader
7. enabled imagetyperz and browser loop

Thats it. Downloads started with the house number captcha and imagetyperz was sent the easier captcha and it was solved and Download worked.

[raztoki]

With linux it has worked fine on all varients I've tested with, with no drama other than if your default browser if firefox.

We are aware with firefox that you need to alter a setting to allow auto close of tabs (its mentioned in the support article). Some of the browser captcha stuff requires tabs to close in order to finish the routine.

What was your default browser before installing chrome?

raztoki

[Jiaz]

Thumbs up for using Linux

Originally Posted by raztoki

With linux it has worked fine on all varients I've tested with, with no drama other than if your default browser if firefox.

We are aware with firefox that you need to alter a setting to allow auto close of tabs (its mentioned in the support article). Some of the browser captcha stuff requires tabs to close in order to finish the routine.

What was your default browser before installing chrome?

raztoki

it was firefox and I was using the beta so probably my install environment that had me running in circles. Plus imagetyperz didnt help by giving me a lot of bad answered captchas when I wasnt paying attention.

chromium seems to work the best for me right now with jdownloader2.

[Jiaz]

Thanks for the feedback!