MarkWilson
Regular Member
- Mar 16, 2014
- 307
- 93
Hey there I made some research regarding this topic as recently I get sign-in blocked by google even with my strongest bot config (browser may not be secure error).
After some hours I encountered this super interesting issue: basically if you are automating chrome, firefox or whatever your are doing so by connecting with remote debugging pipe/port.
In the case of puppeteer, you will find the flag --remote-debugging-pipe in chrome://versions as it is automatically set in order for puppeteer/playwright/selenium to work.
Now it seems Google is able to detect such flag, probably using some javascript object. Any idea which object?
This means that in 100% of cases Google IS ALREADY detecting you, if you're able to pass it's because it let you through, but your account will eventually be flagged and banned or limited later.
My thesis is: no matter how many fingerprints you spoof or how consistent they are, Google will know you are automating just by checking for the presence of the flag.
You can solve this problem by setting or changing User Agent (for example with android UA it works) but I am pretty sure they know this and flag the account.
By the way it seems that starting from January 2021 some bad things will happen for automation
There is so little information about this on the web, some say navigator.webdriver will be set as "true" but that's an old story, there has to be some more.
@Gogol @Sebastiann
After some hours I encountered this super interesting issue: basically if you are automating chrome, firefox or whatever your are doing so by connecting with remote debugging pipe/port.
In the case of puppeteer, you will find the flag --remote-debugging-pipe in chrome://versions as it is automatically set in order for puppeteer/playwright/selenium to work.
Now it seems Google is able to detect such flag, probably using some javascript object. Any idea which object?
This means that in 100% of cases Google IS ALREADY detecting you, if you're able to pass it's because it let you through, but your account will eventually be flagged and banned or limited later.
My thesis is: no matter how many fingerprints you spoof or how consistent they are, Google will know you are automating just by checking for the presence of the flag.
You can solve this problem by setting or changing User Agent (for example with android UA it works) but I am pretty sure they know this and flag the account.
By the way it seems that starting from January 2021 some bad things will happen for automation
There is so little information about this on the web, some say navigator.webdriver will be set as "true" but that's an old story, there has to be some more.
@Gogol @Sebastiann
Last edited: