01
CANVAS & WEBGL
Unique graphics rendering depends on GPU and drivers.
Browser Fingerprint Test
Analyze your browser's unique fingerprint and see how trackable you are
HOW IT WORKS
Complete analysis of your browser uniqueness
02
AUDIO FINGERPRINT
Audio context creates a unique processor signature.
03
FONTS
Installed fonts list is unique to your system.
04
MATH & SYSTEM
Math operations and system parameters.
Your Browser Leaks 40 Facts About You. Here's How Fingerprinting Stitches Them Together.
Before you've clicked anything, a modern browser has already handed over thirty to forty technical facts about itself. Screen resolution. GPU model. The exact font rendering hint your OS applies to Cyrillic ligatures. None of these identify a person alone. Stacked, they form a hash that narrows you down to roughly one device in twelve thousand. Sometimes one in two million. That math powers every "we recognized this device" prompt you've seen since 2014. The stubborn part: these signals can't be cleared. Cookies wipe in one click. Fingerprints don't, because the browser generates them on demand from real hardware. Pull the GPU through a Canvas draw call and out comes a byte sequence specific to one driver build on one machine. Sections below: where that uniqueness actually originates, why private mode does nothing against it, and where fingerprinting's real cracks hide.
Where the Uniqueness Comes From
Browsers expose APIs that look innocent. Read the user agent string. Grab screen dimensions. Ask for a canvas measurement and accidentally reveal which fonts are installed. Each return is a short string with a slightly different frequency distribution across the population. Canvas is the heavyweight. Browser asks the GPU to draw text and shapes onto a hidden surface, then hashes the pixel output. Two machines that look identical on paper produce different hashes because subpixel anti-aliasing depends on the font cache. Font cache state reflects which system fonts are installed, which reflects every piece of software added in the past decade. Somebody's 2019 Photoshop trial still fingerprints their laptop in 2026. Academic research puts Canvas alone at roughly 1-in-1700 identification across mainstream browsers. WebGL adds the GPU identity directly: GL_VENDOR, GL_RENDERER, supported extensions. Strings like "ANGLE (Intel, Mesa Intel UHD Graphics 630)." Specific enough that Chrome on a 2019 ThinkPad with a particular driver pins itself across sessions. Matches nobody else. Audio context is subtler. Run a sine wave through DynamicsCompressor and the float32 buffer diverges between AMD and Intel silicon. The variance sits below human hearing. Well above the threshold where statistics can tell chips apart. Stack these layers and entropy compounds like a centrifuge separating blood fractions: each spin drops the candidate pool from millions to thousands to a handful. Turns out your browser is approximately as unique as your thumbprint, except it volunteers the information to anyone who asks.
What Actually Breaks Fingerprinting
Most defenses fail in ways their users don't expect. uBlock Origin and Privacy Badger block trackers that request fingerprints. The browser still produces them for anyone else who asks because the API is right there. Disabling JavaScript works but craters about 40% of the modern web. Tor Browser takes the standardization route: lock the window size, freeze the User-Agent, route Canvas through a dummy renderer that ignores the actual GPU. Every install looks identical. Individual identification collapses to zero. The cost is a browsing experience that feels like 2008 dial-up, plus banks and CAPTCHA providers blocking Tor exits on general principle. Brave randomizes. Canvas adds per-session noise. WebGL parameters get scrambled. Audio buffers shift by tiny floats. Each session gets a different hash, which kills long-term tracking but lights up anti-fraud systems like a flare. Straight up, it's wearing a ski mask to a bank: face unrecognizable, also immediately suspicious. Anti-detect browsers (Multilogin, Octo, GoLogin) go the other direction entirely. Build a synthetic profile from real device telemetry, hold it stable across sessions, run dozens of these in parallel. At its best, indistinguishable from a real ThinkPad in Cleveland that someone bought five years ago. Cheap implementations leak instantly: their Canvas hash matches no device that has ever existed on Earth. The forgery equivalent of painting a Vermeer on polyester canvas.
The False-Negative Trap
Here's the part nobody warns you about: too clean gets flagged harder than too dirty. Modern anti-fraud runs three checks beyond fingerprint uniqueness. Internal consistency: does the browser claim Chrome-on-Windows but expose Firefox-only APIs? Stability: does the fingerprint change between page loads inside one session? Plausibility: do the reported values describe a device that physically exists somewhere in the world? A randomized Canvas hash trips Plausibility immediately. Real GPUs don't generate fresh entropy on every draw call. They return the same hash for the same input, every time. When yours doesn't, the system writes you down as "automation or anti-detect" and routes you to manual review. Cloudflare's bot scoring actively penalizes excessive randomization. So does Akamai. Some payment processors silently bump fraud scores for any request from a browser that flunks Plausibility, regardless of what the actual transaction looks like. Tor dodges this because uniformity is a coherent defense. Everybody looks the same, nobody stands out. Brave gets caught more than its users realize. Anti-detect browsers walk a tightrope stretched between three cliffs. Profiles need uniqueness or sessions stitch together on next visit. Consistency can't drop or the platform tags them as bots. Aging matters most of all. A profile minted yesterday flags on first contact. Most operators get this catastrophically wrong. The ones who don't charge $100 a month and there's a reason for that. The scan here runs all three signals in parallel and surfaces which one breaks first. Not the headline score. The specific failure mode that would trip a real anti-fraud stack.
Frequently Asked Questions
How can I reduce my fingerprint uniqueness?
Use privacy-focused browsers like Tor or Brave, enable fingerprint protection features, use common screen resolutions, and avoid installing unique fonts or browser extensions.
What does 'Canvas Randomized' mean?
It means your browser or an extension is adding noise to Canvas fingerprinting. This is good for privacy as it makes tracking harder, but some websites may flag this as suspicious.
Is fingerprinting legal?
In most jurisdictions, fingerprinting is legal but may require disclosure under privacy laws like GDPR. Unlike cookies, users cannot easily opt out of fingerprinting.
What is the most unique fingerprint component?
Canvas and WebGL fingerprints are typically the most unique as they depend on your specific GPU, drivers, and rendering implementation. Audio fingerprints are also highly distinctive across different browser engines.
Can I be tracked in incognito mode?
Yes, incognito mode only prevents storing cookies and history locally. Your browser fingerprint remains the same and can be used to track you across sessions. Use Tor Browser for true anonymity.
All our tests
Run the full WHOX toolkit from one place: leak checks, fingerprint probes, network diagnostics, reputation lookups, and privacy failure tests.