Spoof useragent in Tor private tab

The current Tor Browser uses the following useragent string

  • Mozilla/5.0 (Windows NT 6.1; rv:52.0) Gecko/20100101 Firefox/52.0

as Windows users, for every platforms.
I hope your Brave developers team can implement this for better anonymity in your Tor private tab feature.

4 Likes

Yes!! +1 from me.
UA strings are one of the most crucial things to be kept in mind when using tor as they act as a major component in fingerprinting the user & isolating their identity from the whole original majority Tor crowd.

We could change the Brave user-agent to resemble the Tor Browser, but Brave and the Tor Browser behave differently enough that that wouldn’t really improve anonymity — with some effort, Brave is even distinguishable from Chrome, on which it is based, so a completely different browser foundation like Firefox is going to be even easier to tell apart.

What Brave can do is make Brave users hard to tell apart from each other. If you want to blend in with Tor Browser users, the best way is to use the Tor Browser.

2 Likes

@taylor Your claims argue for a Chrome UA string, not a Brave one. The topic is involved. Tor Project barely has a handle on it. Hobnob with tor devs. Please do not think all by yourself.

The OP made a legit request that does work in most cases. UA should be a user setting with best recommended default preset on tor tabs.

If a website has enough JS to scan behavior just to diff Brave from Tor Browser, then a tor surfer probably does not want to see it (except maybe pentesters). The site may not work anyway. On forums even mainstream sites like Chase.com and Craigslist.com have issues. Brave has bigger fish to fry than forcing its UA on tor users.

Full disclosure: Brave has self-interest to promote its UA string. Going by your claims, Brave should not present investors with UA stats. Legit market stats need behavioral or phone-home techniques.

2 Likes

The question of whether the Brave UA string should proclaim Brave or pretend to be some version of Chromium is a separate matter that I’m not personally taking a position on — there are some reasons we might want to look like Chromium, to prevent sites who would block Brave users because of the ad blocker; there are some reasons we might want to promote Brave, to show publishers how many users they can get Brave Payments from. None of this has to do with anonymity among Brave users.

My only point is that if you customize your UA string, you are going to stand out among Brave users and thus it generally hurts your anonymity. We have some open issues about reducing fingerprinting in the default UA string:


“whether … Brave or … Chromium is a separate matter”

Best UA for tor anonymity is the matter at hand. One can’t throw a UA string out sans metrics. That’s waving hands. I hope we’re on the same page in one regard? Brave should transmit different UAs in tor tabs than regular. Maybe you were thinking of a global UA string across all tabs; I’m not.

A tor dev said, “The Tor Browser has a unique fingerprint, this is unavoidable as long as other browsers don’t care about privacy issues. But ideally it should be the very same on every different computers it is used so that we all belong in one anonymity set.”
https://trac.torproject.org/projects/tor/ticket/9501

Underscore: ONE ANONYMITY SET.

“None of this has to do with anonymity among Brave users.”

That’s ANOTHER SET. It’s erroneous to conflate the two.

“My only point is … you are going to stand out among Brave users and thus it generally hurts your anonymity.”

No. The point of tor is to POOL TOR USERS. Making Brave users indistinguishable among themselves does not help them on tor. They need to pool with tor users, not Brave users.

By your claims we could just as well use a ‘Brave Anonynet’ apart from tor. Why? The Brave subset on tor is easy to nail.

Think numbers. Tor slows browsing. Only those keen will use it, not the whole Brave user base. So if (say) only 30 Braves use tor at once, among (say) 10000 generic tor users, the Brave folk stand out like neon signs. Just by UA string, no other reason.

Quibbling behavior argues naught against a minimal defense UA, trivial to implement. Full UA customization is debatable, not a choice based on pooling stats. Deeper fingerprinting just means Brave has more work to support tor. Don’t make the perfect the enemy of the best. Ask tor project advice.

It would be nice for anonymity if the Tor Browser and Brave could behave identically from an adversary’s perspective. But as long as Brave aggressively blocks ads as part of its basic functionality and the Tor Browser doesn’t, that would be hard to pull off. (And of course Chromium, on which Brave is based, and Firefox, on which the Tor Browser is based, are going to behave differently in various ways due to the completely different browser architecture.)

Similarly, it would be nice for anonymity if Chromium and Brave could behave identically from the web server’s perspective. But Chromium users don’t normally use the Tor network, and don’t normally block ads or trackers or fingerprinting vectors either, so those alone will distinguish Chromium users from Brave users anyway.

So the best that it is practical for us to achieve is to make Brave users less distinguishable from one another. It would be nice if all users were indistinguishable, but that’s not achievable today.

Given that, the choice of which user-agent string we use doesn’t make much difference to anonymity as long as everyone has the same one. For defence against very casual attackers, we might choose a Chrome-like user-agent, or a Tor Browser-like user-agent. But there are other issues about what goes into the user-agent string like web compatibility that aren’t germane to this discussion.

Ahem Ahem!

https://community.brave.com/t/a-controversial-topic/31352

The solution is very simple: let the advanced users decide for themselves. As stated above, let the user set the UA. Problem is instantly solved.

Sorry to say, Brave has always been too rigid in deciding everything for all users. That’s why I stopped using Brave a few years ago. Brave allows google to spy on everybody in non-TOR tabs.

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.