Browser Fingerprint: Am I Unique, and How Do I Check?

What “Am I Unique?” Really Asks About Your Browser

Probably yes — you are unique, or close to it. When you ask browser fingerprint am I unique, the real question is how much identifying information your browser hands over while a page loads. A test reads signals like canvas rendering, WebGL output, installed fonts, screen size, timezone and language, then compares that combination against a stored dataset. The fewer browsers that match yours, the more unique — and the more identifiable — you are.

How a Fingerprint Test Decides You’re Unique

Fingerprinting runs while the page loads. It needs no permission prompt, writes nothing to disk, and survives clearing cookies or opening a private window. That is the difference from cookie tracking: there is nothing to delete.

Tests report the result one of two ways. A similarity or uniqueness percentage comes from matching your values against a stored dataset. Entropy, measured in bits, describes how much separating power a value has — each bit roughly doubles the number of distinct browsers it can split apart. A value that separates one browser in four is worth 2 bits; one that separates one in a million is about 20. So the answer to browser fingerprint am I unique depends on which number you read: a percentage describes your whole configuration, entropy scores one signal at a time.

For scale, an INRIA study of roughly two million fingerprints, cited by AmiUnique.io, found 35.7% of desktop browsers were unique and 18.5% of mobile browsers were unique. The same source argues fingerprinting is getting more common as cookies become less reliable, and notes that in 2024 Google said it would no longer prohibit its advertising customers from fingerprinting — a change the UK ICO criticised.

AmIUnique’s own results page shows how much the headline depends on which attribute you read. A full desktop Chrome user agent scored 0.00% similarity, meaning nothing else in the dataset matched it, while the content-encoding string gzip, deflate, br scored 53.92%. Same browser, opposite verdicts. The arithmetic behind those numbers is walked through in fingerprint entropy explained.

Population tests versus mechanism tests

Population tests — EFF Cover Your Tracks, AmIUnique — tell you how many other browsers look like yours. Mechanism tests, such as ReadMyBrowser’s per-signal page, estimate what each individual signal is worth and deliberately use no population dataset. The first answers “am I rare?”, the second answers “why am I rare?”. You want both readings before you change a setting.

What the number does not tell you

It says nothing about whether you are being tracked today or how a site will behave for you tomorrow. A test is a snapshot of how unusual your configuration is, measured against one tool’s sample of the web.

Test Your Browser Fingerprint in Seven Steps

Work in order. The goal is a record you can diff later, not a screenshot you never reopen.

1. Read the method before you press the button. Cover Your Tracks compares you against a database of 5M+ fingerprints and reports bits of identifying information. AmIUnique shows a similarity percentage per attribute against a research dataset whose no-JS page cited 5,614,805 fingerprints. Pick based on whether you want a verdict or a breakdown.
2. Open the exact profile you care about. Fingerprints are scoped to the browser profile, not the machine. A profile with different extensions, fonts or window size returns a different result, so a fresh default profile tells you nothing about the one you actually use.
3. Run the first scan untouched. No resizing, no toggling a VPN mid-scan, no installing extensions with the page open. AmiUnique.io says its scan runs only on a click and takes about five seconds; leave the window alone until it finishes.
4. Write down raw values, not the badge. Tool, date, percentage or bit score, plus canvas hash, WebGL vendor and renderer, font count, audio hash, screen resolution, timezone and language.
5. Expand the full attribute list. Find the attributes where you are the only entry in the dataset. Those are worth fixing; the rest is noise.
6. Cross-check with a tool that uses the other method. If two tools agree on fonts and disagree on canvas, you have learned something about the tools as well as about yourself. Compare signals, never headline scores across tools.
7. Change one setting, quit the browser fully, re-run the same test. Two changes at once and you cannot tell which one moved the number, or whether one signal compensated for another.

How to Read Your Result Without Misreading It

Read every score relative to the dataset behind it. A percentile against 3,288 fingerprints is a much weaker claim than one against five million, even though both produce a dramatic-sounding verdict. The honest answer to browser fingerprint am I unique is therefore a range tied to one sample, not a fixed property of your machine.

Note what each tool stores, not just what it shows. AmIUnique sets a four-month cookie when you contribute your fingerprint; WhereAmI.com’s rolling 30-day corpus keeps only a hash and the day. Read the page first, and if you would rather not contribute, test in a profile you do not use for signed-in work. For a wider look at how these pages differ in depth and reliability, this comparison of browser fingerprint test tools compared covers the trade-offs.

Which Signals Do Most of the Work

Not all attributes are equal. On ReadMyBrowser’s mechanism page, the WebGL renderer and canvas rendering together contribute more than all other signals combined. Older research points the same way: Tor Project documentation from 2018 called HTML5 canvas the single largest fingerprinting threat browsers face once plugins and plugin-provided information are set aside.

Canvas

Canvas fingerprinting draws text and shapes into an invisible HTML5 canvas element and reads the pixels back. The output varies with the GPU and graphics driver; text rendering varies with installed fonts and anti-aliasing. BrowserLeaks extracts pixels with toDataURL(), encodes them as base64 and hashes with MD5 to produce the signature it displays; the CRC in the PNG’s IDAT chunk works as an alternative. The technique became widely known in 2014 when Princeton and KU Leuven researchers published The Web Never Forgets — AddThis had already been experimenting with it that year, and 5% of the top 100,000 sites used it while deployed. The full extraction path is covered in this guide to canvas fingerprinting explained.

How far does canvas narrow things down on its own? In a small study of 294 Mechanical Turk participants it produced a measured 5.7 bits of entropy, and the authors expected more in the wild — modest alone, which is why it usually reinforces other signals. DrawnApart, published in 2022, extended fingerprinting to graphics hardware performance characteristics and raised tracking duration by 67% when used to enhance existing methods. More background sits on the canvas fingerprinting reference page.

WebGL

The renderer and vendor strings name your graphics stack, often specifically enough to identify a GPU model and driver family. Paired with a canvas hash, that combination does most of the work of separating you from the crowd.

Fonts, audio and hardware

Font lists are long, poorly standardised and highly variable between machines. Audio fingerprinting measures tiny differences in how the audio stack processes a signal. Screen resolution, colour depth, device memory and CPU core count are low-value individually but free to collect and easy to combine.

Signals that follow your network

Timezone, locale, geolocation and WebRTC candidates follow your connection. An IP in Helsinki with a Los Angeles timezone is not a fingerprint value in itself — it is a consistency flag that stands out in any review of the session.

Settings That Actually Lower Your Identifiability

Most advice on this topic is a list of features to switch off. Switching features off one at a time usually makes you rarer, not less identifiable. These six moves behave differently, because each one attacks the real question behind browser fingerprint am I unique: how large is the group that shares your configuration.

1. Use a browser that standardises signals for everyone. Tor Browser makes all its users look identical to one another, which is the only approach ReadMyBrowser says fully works. The cost is speed and site compatibility, and it is not a practical daily driver for most work.
2. Or turn on built-in fingerprinting protection. EFF recommends a tracker blocker such as Privacy Badger, or a browser with fingerprinting protection built in. In Firefox you can set privacy.resistFingerprinting in about:config — WhereAmI.com calls it maximally protective but warns it breaks some sites, so test the ones you need first.
3. Stop swapping your user agent. Replacing a user agent shared with millions for a rare one makes you easier to single out, not harder. It is the most common self-inflicted wound on this list.
4. Choose consistency over randomness. A value that changes every session is itself a tracking signal, because the change is observable. Pick a plausible, common configuration and keep it stable.
5. Do not simply disable WebGL. WhereAmI.com notes it costs some 3D content, and ReadMyBrowser calls removing one signal worse than nothing if the result leaves you in a smaller group than before — a browser with WebGL off is an unusual browser.
6. Separate the contexts you legitimately manage. If you handle several accounts you own or are authorised to manage, give each its own browser profile with its own cookies, storage and proxy. A profile manager such as Send.win applies this per profile, so each keeps a stable set of signals instead of mixing them together.

For teams running many profiles at once, the harder problem is keeping each identity internally consistent — proxy exit IP matching timezone, locale and geolocation, with no leakage between profiles. This guide to browser fingerprinting isolation covers how to structure that. Where a platform limits multiple accounts, use its official route — Amazon’s approval process for an additional seller account, or Business Manager roles for agencies working on Meta — rather than adding profiles the platform has not authorised.

Pitfalls That Make You More Unique

These moves reliably backfire, usually because they replace a common value with a rare one.

• Swapping to a rare user agent. A user agent shared with millions is camouflage; a distinctive one is a label. Changing it is the most common way people make themselves easier to pick out.
• Blocking canvas outright. Tor Browser notifies you of canvas read attempts and can return blank image data, but it cannot tell a legitimate use from a fingerprinting attempt. You break image editors and charting libraries, and you look unusual doing it.
• Randomised or mismatched values. A screen size that does not match your real viewport, or a timezone that does not match your proxy’s exit IP, is a consistency flag. It produces a combination nobody else has.
• Assuming a VPN helps. A VPN changes your IP address and nothing else. Canvas, WebGL, fonts and audio come out exactly the same.
• Assuming private browsing helps. Incognito clears local state. It does not change a single fingerprint signal.
• Changing settings in the wrong profile. Fixing your default browser and then testing the profile you actually use is the most common reason people conclude “the fix didn’t work”.
• Trusting a vendor percentile. AmiUnique.io calls itself spy resistant and runs lie-detection checks; that is a vendor claim, not an independently verified result. Read the methodology before you read the score.

How Send.win Helps With Browser Fingerprint Am I Unique

Send.win is a browser profile manager for professionals and teams: every account gets its own isolated browser profile.

• Isolated profiles – separate cookies, storage and sessions per profile
• Fingerprint protection – each profile keeps its own consistent browser fingerprint, so profiles never share one browser identity
• Desktop app + cloud browser – native app for Windows, macOS and Linux, or run profiles in the cloud with no install
• Residential proxies included – on Pro and Team, with timezone and locale matched to the proxy
• Team sharing – share a signed-in profile with a teammate without sharing the password

Try the free cloud demo — no install — or download the desktop app. Paid plans start at $19/month, and your first card purchase is covered by a 7-day money-back guarantee (see pricing).

Verify the Fix: Re-Test the Same Way

Verification is the step almost everyone skips. Run the same tool you started with, at roughly the same window size, in the same profile, with the same proxy connected. Compare individual attribute values first and the headline percentage second — a canvas hash that changed tells you more than a percentage that moved two points.

Then run the second tool. If both now place you in a large group on the signals that used to single you out, the change worked. If one signal improved and another got worse, you have traded one rare value for another and should revert. Repeat on a second day, because extension updates and font installs move values without your involvement.

If you manage several profiles, test each one individually. A shared profile list can look consistent from the outside while one profile inherits a proxy setting or extension the others do not have. The only reliable check is opening each profile and running the same test inside it.

Frequently Asked Questions

Am I unique? How do I check my browser fingerprint?

Open EFF Cover Your Tracks or AmIUnique in the exact profile you use and run the test once without changing anything. Both return a uniqueness verdict plus a per-attribute breakdown. Record the raw values — canvas hash, WebGL renderer, fonts, timezone — because those are what you compare after making changes. That first run is the only honest answer to browser fingerprint am I unique for that profile.

Does incognito mode hide my browser fingerprint?

No. Private browsing clears local state such as cookies and history when you close the window. Fingerprinting reads signals from the browser as the page loads, needs no storage and no permission, and returns the same values in a private window as in a normal one.

Does a VPN hide my fingerprint?

It changes your IP address and nothing more. Canvas rendering, WebGL output, installed fonts and audio processing stay identical. A VPN does help when your timezone and locale must match your exit IP, but that is a separate problem from how unusual your browser looks.

What is a good uniqueness percentage on AmIUnique?

Lower similarity means more unique, so there is no target score to chase in the abstract — the useful comparison is before and after a change, measured on the same tool. A full desktop Chrome user agent scored 0.00% similarity on AmIUnique’s own results page, meaning nothing in the dataset matched it.

Is canvas or WebGL the bigger fingerprint signal?

They are close and they reinforce each other. ReadMyBrowser reports that the WebGL renderer and canvas rendering together contribute more than all other signals combined. Canvas captures GPU and driver variation through pixel output; the WebGL renderer string names the graphics stack directly.

Are fingerprint tests safe to use?

Reading your own result is harmless. Contributing your fingerprint to a dataset is a choice — AmIUnique sets a four-month cookie when you contribute, while WhereAmI.com’s rolling 30-day corpus stores only a hash and the day. If you would rather not contribute, test in a profile you do not use for signed-in work.

How do I make my browser less unique?

Blend in rather than stand out. Turn on built-in fingerprinting protection instead of disabling features one by one, stop overriding your user agent, keep one configuration stable across sessions, and keep separate work contexts in separate profiles. Re-test after each change to see whether the number actually improved.

Is AmIUnique still updated?

Treat its dataset figures as historical: the no-JS results page cited 5,614,805 fingerprints, and parts of the project carry dates from several years ago. Newer independent tests have appeared since, including AmiUnique.io with a 3,288-fingerprint corpus, so any single dataset is a reference point rather than a live census of the web.

Related Products & Resources

• Download
• Demo Browser
• Pricing
• Fingerprint Checker
• Canvas Fingerprint Test