[STUDY] Proxy 4g fingerprint and possible reasons for the last blocks on automations

Hello everyone, during this time I did a lot of tests on my 4g proxies, I read a lot, and heard opinion of several people.
So I decided to write a small study containing some of the key points that I think are generating many problems for those who use automation.

These are the results of the 4g proxy tests and some other considerations.

To begin with I analyzed the fingerprints (outside JV) with sites like www.f.vision
By analyzing the IPs masked by proxies, these sites are able to detect the real IP address to which we are connected, this via WebRTC.

So the sites detect two IP addresses, the first of the proxy and the second (local IP address) which is the real IP address to which we are connected.
Obviously it is very simple to disable WebRTC with a browser extension, but obviously sites like f.vision detect an inconsistency (the result is FAKE WEBRTC DETECTED) given by the fact that the WebRTC has been deactivated.

By running tests (inside JV), with embed browser, I note that many types of fingerprints have been hidden, but unfortunately not all of them.

I noticed that for example JV hides a lot of information to protect the identity of the user, it is certainly a positive thing, but at the same time instagram will detect that there is a difference with respect to an average normal user, the average user does not use the proxy, does not hide your data, does not hide anything.

Analyzing with whatleaks.com (in JV in embed browser) a difference between operating systems is detected.
JV fake user agent is android, but browser windows requests.

Analyzing instead from a simple smartphone is android / android.
Even privacy.net has interesting fingerprinting analysis methods, but the most frequently detected problem is that it is the double operating system.

Another test on the passive os fingerprint, if the proxy is done well (or done with a smartphone) the fingerprint is correct (linux), if you use a hotspot it results Windows.

Again, if you analyze the passive os of a simple smartphone it is linux / linux.

Another thing detected is the geolocation (really important in my opinion).
Using proxy 4g of my city the geolocation is perfect, using proxy of another state instead no, this can make you think.

Another thing to think about, in which I didn’t do tests, but logically it seems to me absolutely plausible for those who use proxy 4g: The rotation times of the IP address.

Do you think it is normal that an IP address changes perfectly every 3 minutes or every 5 or every 60?

I don’t think so.

When we are outside home it is unrealistic that the phone goes offline or we change position with a precise time interval.

Now imagine dozens or hundreds of accounts, which change IP address with perfect frequency of 3 minutes, all on the same geolocation, network, similar ip, same operator.
In my opinion, an extremely visible pattern is created.

Moreover, as if this were not enough, groups of 10 account each time fall on a different IP address, but the same is the same for all these 10 accounts, and then they change IP again, and then return to another IP again equal for all 10 accounts, all this repeated dozens of times a day.

It seems obvious to me that an extremely visible pattern is created.
However it would be much more realistic that the IP address changed at random intervals, and that the accounts rotated between the various proxies/ IP, and the same 10 accounts did not always remain on a same IP address.

Another fundamental element is to make the actions of the accounts as random as possible, the people are unpredictable, the humans are chaotic, it is absurd in my opinion that a person goes 12 hours a day every day always in the same time bands and every hour it always puts the same number of likes.

The same goes for those who perform actions by hand, it is useless to do them by hand on the official app if we then move exactly like a robot, with patterns that are always the same.
Even on the app you have to be human, add in ever-changing times and numbers.

Another point on which we are talking about is the fact that the current automation tools are not able to perfectly emulate human actions such as mouse movement or touch on the screen.

Another point is also the ability to replicate all the information given by one smartphones, such as battery power, camera, processor and many many others.

I hope this study can be useful to all of you.

Greetings to all.


Useful read about IPs/Proxies. But manually liking/following has no problems if you follow the limits. Tested & Viable. My opinion :wink:

1 Like

Thanks, surely by hand it is the best choice if you have few accounts/clients to manage.
In case you have many accounts to run by hand it can become a problem.

1 Like

@reverse Thank you for this detailed post!

1 Like

You are welcome. :wink:

1 Like

Interesting take on Geolocation, I may try faking it on my next project to match random coordinates near the city the proxy is located in, see if it increases success rates.

Thats true, but still you have options to hire people to follow/like your stuff by hand.

1 Like

Thanks, anyway I don’t know if faking geolocation can increase success rate, I would be more secure going to a 4g proxy in my city, or creating proxies with phones from home.

1 Like

JV leaving some footprint is a strong suspect, even if these footprints come from hiding too many things. The Dev Team has confirmed working on it, and I assume your excellent study will help them here.

About the proxies… We have been using all sort of proxies, and for now, it doesn’t make a huge difference but it helps to have good ones.

I don’t think that location matters that much. Our clients have different members of their team logged in the account, from different location and we are using proxies from another location. It doesn’t bring us blocks on these accounts.

Excellent job mate.


Thats exactly what I was talking about JARVEE.

It doest generate RANDOM HASHS. IT JUST BLOCK!

1 Like

I tried to write to JV’s help by sending them my research, but I didn’t get an answer. If you want you can try sending it too, maybe you’ll be luckier.
Months ago I worked with proxy datacenter, but I had continuous blockages on folow, and I couldn’t put like, never. Now with my 4g proxies I no longer have these blocks.
What I mean about geolocation is that the VPS / JV / browser coincides with the proxy location.
Anyway, thank you for reading carefully my study.

1 Like

we are working with agencies to handle their accounts manually doing 200 follow, 200 unfollow, 150-300 likes daily with 0 problems on 55+ accounts, so yes if you follow the limits you will not get any blocks while doing actions manually

1 Like

kudos to you, cannot imagine doing that amount of interactions by hand every single day on 50+ accounts

lol im not doing those actions myself i got people :wink:

1 Like

obviously, but thats only stress/pay above

in theory jv should be doing that

In theory J shouldn’t exists since its against their TOS. This is a social media marketing forum, not How-To J forum.

agreed yet you can see most of topics in “Instagram marketing” category revolve around how-to J/M
at least in past couple months since big issues emerged :man_shrugging:

Very good points explained here. I try to tell people that Jarvee is doing shit hiding instead of faking it, but noone listen to me.

1 Like

If you could let me know, that would be great :slight_smile: