Hardened browsers compared. Tor, Mullvad Browser, Brave, and Firefox.

Two defense strategies

Every anti-fingerprinting browser picks one of two strategies. The first is uniformity: make every user of the browser look identical, so the fingerprint identifies the browser, not the person. Tor Browser and Mullvad Browser do this. Same window dimensions, same fonts, same user agent string, same answers to every probe. You hide in a crowd, and the crowd has to be large enough to matter.

The second is randomization: feed trackers a slightly different fingerprint on every site or session, so the data points never connect into one profile. Brave does this, the project calls it farbling. The signals look normal, they are just subtly wrong, and wrong differently each time.

Both strategies work. They fail differently. Uniformity breaks the moment you resize the window or install an extension. Randomization survives sloppiness better but can trip fraud detection on banks and payment platforms. The mechanics behind each signal are covered in our guide to browser fingerprinting in 2026.

The four contenders

Tor Browser

The reference implementation of uniformity, and the only one that also hides your IP by design through the Tor network. Letterboxing locks window dimensions to fixed steps. Canvas, WebGL, audio and font probes return the same values for every user at the same security level. The crowd is in the millions.

The cost is operational. Exit relays are blocked or flagged by a large share of commercial sites. Logins trigger verification loops. Speed is a fraction of a direct connection. Tor Browser is a tool you open for a task that requires anonymity, not a place to live. Using it for your named accounts defeats the point and contaminates the session.

Mullvad Browser

Built by the Tor Project with Mullvad, it ships the Tor Browser hardening without the Tor network. You get the uniform fingerprint, letterboxing included, over your own VPN connection. Speed is normal. Sites treat you like a VPN user, not a Tor exit.

The trade is explicit: no Tor network means your VPN provider sees your traffic, so the browser is only as anonymous as the tunnel under it. The documented failure cases are collected in you can be tracked through a VPN. The crowd is also smaller than Tor Browser’s, which weakens the uniformity guarantee at the margins. For sustained pseudonymous work over a trusted VPN, it is the strongest daily option.

Brave

The only Chromium option in the list, and the only one a non-technical person can run as a primary browser without breakage. Farbling randomizes canvas, WebGL and audio readouts per site and per session. Built-in blocking removes most trackers before they probe anything. Standard mode covers most users; Strict mode tightens the randomization at the cost of more site breakage.

Limits: the fingerprint surface of Chromium is large, randomization covers the noisiest probes but not all of them, and a motivated adversary correlating IP plus behavior still wins. Brave reduces commercial tracking dramatically. It is not an anonymity tool.

Firefox with resistFingerprinting

The hardening that Tor Browser is built from, available in stock Firefox behind the privacy.resistFingerprinting flag, usually deployed with the arkenfox user.js template. Timezone reports UTC, canvas prompts for permission, window sizing is stepped.

This is the power user route. Nobody maintains the configuration for you, an update can silently change behavior, and the crowd of arkenfox users is small enough that the configuration itself becomes identifying. Run it because you want control and understand the trade, not because it is stronger than the alternatives.

Test before you choose

Opinions about fingerprinting are cheap. Measurements are not. Run your current browser through our Fingerprint Analyzer, note the score, then run the candidate browser on the same machine. The 22 signals it reads, and what stops each one, are documented in the operator guide to website tracking. A browser that drops your exposure score by half changes your situation. A browser that moves it two points does not.

Verdict by profile

Contacting a source, researching an adversary, or doing anything where the session must not link to you: Tor Browser, full stop. Accept the friction, it is the price of the crowd.

Sustained high-risk work under a real name or stable pseudonym, journalist drafting, NGO reporting, legal research: Mullvad Browser over a paid, audited VPN. Uniform fingerprint, usable speed.

A civilian or family member who needs one browser for everything, banking included: Brave in Standard mode. It is the only option here that survives contact with normal life.

Compartmentalize regardless of choice. One hardened browser for sensitive work, one ordinary browser for named accounts, never mixed. Which compartments you actually need is an output of your threat model, and you can build one in twenty minutes with our guide on how to build a threat model.

Living with a hardened browser

The choice you make today decays. Browser updates change fingerprinting surfaces, uniformity crowds shift as users migrate, and the trackers update faster than any of it. The hardened browsers above handle the maintenance for you, which is most of their value: Tor and Mullvad Browser ship hardening as defaults that survive updates, Brave retunes farbling continuously. The configurations you assemble yourself, arkenfox included, are the ones that silently drift.

Workflow matters more than settings. Keep the hardened browser boring: default window size, no extensions, no logins that carry your name. The moment a pseudonymous session and a named session share a browser, a cookie jar, or a fingerprint, the compartment is gone and no setting brings it back.

And re-test on a schedule. A quarterly pass through the analyzer takes five minutes and catches both the browser regressions and your own drift. Treat the score as a metric you maintain, not a box you ticked once.

Frequently asked questions

Is Safari acceptable for privacy?

Safari ships real tracker blocking and some fingerprinting defense, and Apple hardware uniformity helps. But configuration options are minimal, the protections are opaque, and on macOS the fingerprint still exposes the exact hardware tier. Fine as a civilian default, not a hardened choice.

Do extensions improve or weaken these browsers?

They weaken them. Every extension changes behavior that probes can measure, and a rare extension combination is itself a fingerprint. Tor and Mullvad Browser are designed to run as shipped. On Brave, the built-in blocking already covers what most privacy extensions do.

Does a VPN make fingerprinting protection unnecessary?

No. A VPN replaces your IP address and nothing else. Canvas, fonts, screen profile and the other stable signals pass through the tunnel untouched, which is exactly how users get re-identified across VPN sessions.

What about mobile?

On Android, Tor Browser and Brave both exist and behave like their desktop versions, and Mullvad Browser does not. On iOS every browser is WebKit underneath, so the engine-level defenses are capped at what Apple ships. Mobile hardware is also more uniform, which helps, and more personal, which does not.

---

There’s no perfect setup. Anyone selling you perfect is selling fear. The goal is simple: make yourself a harder target than the person next to you.