First of all thank you very much for taking the time to explain all this in detail, it is instructive for some, and it triggers a deeper analysis of the problem.
As support mentioned you should know that reCaptcha and Google services registrations does not use the same techniques to determine the authority of the visitors.
I do not know if for sure i understand what you mean by authority of the visitors, to clarify,
Yes that's right, the problem here is that we're talking about visitor "classification", reCaptcha is not to determine good users or bad users, it's to determine "human" users OR "bot" users. Captcha means: Automated Public Turing test to tell Computers and Humans Apart.
i understand of course with these known technique that you explain is used by new detection system to identify bots, again this is not the case, in my opinion Recaptcha does not use native function detection, if a function like canShare() (or any other function, even CSS) is called in a browser which does not support this function we find in the console a nice
Uncaught ReferenceError: canShare is not defined
We deduce that Recaptcha detects something else and not the native nonexistent/existing browser function. (i assume that you wase talking about the native function of the BROWSER)
There are function compatibility databases like caniuse.com, we can see that the function you mentioned above is chrome 81 comptaible on android : https://caniuse.com/#search=canShare
I mean that is possible to obtain a list of missing functions by comparing two versions of chrome, it would be wunderful for bablosoft to code a plugin to synchronize the fingerprint to add/remove (or simulate ...) the necessary functions, it is only an idea ... which would surely reduce the risk of function detection, not completely of course because it is not enough to return a true or false to totally be sure that the native function are operational or not, it will just be operational to answer false or true, that's good for a start. example by comparing chorme 70 and chorme 81: https://caniuse.com/#compare=chrome+70,chrome+86&compare_cats=JS,JS API
we get the difference function compatibility between the two browsers ...
If wee deepen the analysis, what can happen other than the detection of native function or verification of fingerprint/canvas ... what remains ?
A extrapolate theory would be the Fingerprint of the TCP/IP stack in order to deduct the OS and to compare it to the OS send by header, i strongly doubt that this is the case, firstly because of the execution time of this process which is a little long ... in the doubt, i made a config in perfect correlation with my system, in order to test in the most optimal environment to counter the two verificaiton evoked (native function/OS detection):
My real OS is windows 10x64.
The config:
All right, screenshot here:
We Check now ths Recaptcha score :
Score 0.3
The test Project : NativeFunctionAndOsBypass.xml
So, can we now deduct that is not about native function and/or OS détection ?
With all this said BAS is the most throughoutly implemented framework for automation whatever yo uuse it for specially the fingerprint technology. And most sites you work with you won't have a problem as most sites doesn't have advanced detections like this. You can try compare Geckodriver with Chromedriver yourself if you want to explore more between the difference automating Firefox and Chrome.
After having already try all ..., it is undeniable, BAS is great, with a high level of development.
On the other hand i can provide screenshoot or video of anti-detection browser which pass Recaptcha V3 whit a score of 0.9, a f.vision whit score of 100% ... (Geckodriver), and this with fingerprint/user-agents which are different from my OS, so there is always possible improvements no matter the level reached, i'm still looking for a solution to raise the score at least 0.7 ... i would post here if I find something ...
Thank you all, Thank you again @0xSentinel
No these example ways how to detect browser versions from other element except Canvas, user-agent header etc.
TCP stack sounds overkill I dont think google use it, maybe mostly for banks and financial institutions.
Haven't tried recaptchav3 much since few months back where I scored 0.7 many times but most times were 0.1 and 0.3 using fingerprint switcher. I can look into it again when I have time later. For now I don't think problem because not many sites uses recaptcha v3 at the moment but as soon as it becomes a problem to me I will make sure to find a solution.
Have you tried set user agent to as close to stock Chromium as possible? I think this is what CEF is based on so using Googles own product as useragent and hope to bypass their own system is hard job. Better try set the user agent header for Chromium, Brave, etc and other browsers using Chromium.
Yes it is exactly true, my bot operates in the following way :
I take for example my works from this last month, the bot has never exceeded 3 or 4 attempts before having a fingerprint which obtains a score of 0.7 at least, today i can let it run for hours it no longer finds any fingerprint which get socre more than 0.3, so i wase thinking of a problem with fingerprints from fingerprinswitcher, to test this hipothese i use a profile that have a fingerprint which has a score of 0.7 on Recapctah V3 befor the problem, and it does not pass the test, it now gets a 0.3, it seems that the problem does not come from fingerprintswitcher.
I can look into it again when I have time later. For now I don't think problem because not many sites uses recaptcha v3 at the moment but as soon as it becomes a problem to me i will make sure to find a solution.
It is not my case :) I am impacted by this problem and must find a solution to be able to continue to use BAS :)
TCP stack sounds overkill I dont think google use it, maybe mostly for banks and financial institutions.
Yes, as i said i strongly doubt that this is the case for google, it was just to evoke all the possibilities on this post in order to avoid future quesitons...
Have you tried set user agent to as close to stock Chromium as possible ? I think this is what CEF is based on so using Googles own product as useragent and hope to bypass their own system is hard job. Better try set the user agent header for Chromium, Brave, etc and other browsers using Chromium.
You are right It is quite obvious, i will try and give you my feedback.
thanks
@jonigood How are you even getting an score of 0.3? I am getting only a score of 0.1 which causes me to get a lot of captchas, How are you getting a score of 0.3? Also how are you putting in verification checks to get fingerprints that are above a specific score? Could you help me do that?
@tatwa Hi, each user of BAS have these tips and configurations to get a better score, it depends on a lot of parameters, you must configure and test your project before launching the recaptcha test, you must pass all the online detection tests like f.vision ect ect ... this n 'is that you have to try recatcha, i provided one or two test project on this post, try them ...
@jonigood I am passing all the online detection tests and that is not what I am asking about, What I am asking is how are you implementing a test to filter out fingerprints that give a low score and only select fingerprints that are high quality and use them?
I scrolled through with my regular browser with regular activity. Recaptcha must have had an update as I was hit with lots of captchas. Before I got about 80-90% of the times green checkbox but the last day I have had to solve fast fading images.
This is with a stock browser and clean IP.
No wonder then it's harder to get good score.
Figured I tell you this, may be valueable to you.
@FastSpace Я раньше рукапчей увлекался, так там после 10-15 минут решения рекапчи в боте вообще score до 0.1 падал. Кроме прогретого профиля влияет и количество решённых за последнее время рекапч, и то, насколько давно вы последний раз грели профиль и, наверное, уйма других факторов.
friends, please write in english, we don't understand notthing, google translat is ineffective with abbreviated russian language ...
@FastSpace Что ты какой-то сказочник. Скор дает еще гретый профиль.
Кстате у меня с кучей авторизации в нормальном хроме тоже 0.3.
I never had less than 0.9 on a real chrome
@0xSentinel hi, Recaptcha V3 updated its code betwen the 27/06/2020 and 28/06/2020 ... it is now very "difficult" (to not write "impossible") to obtain an acceptable score of 0.7 whit BAS.
