AmIUnique is an academic fingerprinting research site. Learn what its uniqueness result measures, how its dataset skews it, and what it can't tell you.
AmIUnique is a different kind of fingerprint test. Most tools people run are commercial dashboards or advocacy demos; AmIUnique describes itself as a research project studying the diversity of browser fingerprints, with the aim of giving developers data to help them design good defenses. Its tagline is "Learn how identifiable you are on the Internet", and the website's source code is published openly in the AmIUnique repository on GitHub. That makes it a neutral tool rather than a product trying to sell you something, but it doesn't make its headline result a verdict on your privacy.
Key Takeaways
- AmIUnique is an academic research project from Inria, not a vendor: the site's own description is that it studies fingerprint diversity, and its source is public.
- It reads HTTP headers plus attributes collected by JavaScript, then compares your combination against the fingerprints it has stored itself.
- "Unique" means unique within AmIUnique's own dataset, which is not a random sample of the web. Visitors tend to be people curious about privacy.
- A unique result is about trackability. It says nothing about whether your browser is spoofed, a bot, or behind a VPN.
- Its distinctive angle is looking at fingerprints as something that changes over time, not as a single snapshot.
Where AmIUnique Comes From
AmIUnique was launched in 2014 by researchers at Inria, the French national research institute for digital science, and its data has fed peer-reviewed papers rather than a product roadmap. The best-known is "Beauty and the Beast: Diverting modern web browsers to build unique browser fingerprints", presented at the IEEE Symposium on Security and Privacy in 2016, which analysed fingerprints collected through the site, including from mobile devices. A follow-up at the same conference in 2018, "FP-STALKER: Tracking Browser Fingerprint Evolutions", used fingerprints gathered repeatedly through AmIUnique's browser extensions to study how fingerprints drift and whether they can still be linked. That lineage is the reason to read its results carefully: they are research measurements, with the strengths and the limits research measurements have.
What AmIUnique Actually Measures
When you run the test, your browser sends the usual HTTP request headers (the user-agent, accepted content types, language and encoding preferences), and a script on the page reads a set of JavaScript-visible attributes: things like screen geometry, time zone, plugins, and rendering output from canvas and WebGL. The server stores that combination as a fingerprint and counts how many fingerprints already in its collection match yours, attribute by attribute and as a whole.
For the exact, current attribute list, read the site's own explanation and the published source rather than a blog summary, including this one: collectors get added and retired as browsers change, and a stale list is worse than none. The signal families themselves are the same ones every browser-fingerprinting tool relies on. If you want each family explained one by one, BrowserInsight's fingerprint check shows the same kinds of signals with a short explanation beside each.
Reading "You Are Unique" Correctly
The result is a ratio against a database, so the denominator decides everything.
The dataset is self-selected. People who find a fingerprinting research site are, on average, more privacy-aware, more technical, and more likely to run unusual setups (extra extensions, hardened browsers, uncommon operating systems) than the general public. A browser that looks rare in that crowd can be perfectly ordinary on the wider web, and the opposite can also happen. The same dataset-relative caveat applies to EFF's tool, which we covered in Cover Your Tracks: What Your Uniqueness Score Doesn't Prove.
Rare is not suspicious. A stock browser with an uncommon screen size or a rarely installed font can be unique and entirely genuine. Uniqueness measures how easily you could be recognised again; it is not a coherence check. Catching contradictions between signals, like a user-agent that claims one browser while the rendering output says another, is a different job, covered in CreepJS Explained: How Fingerprint Lie-Detection Catches Spoofing.
The numbers don't transfer. A figure from one site isn't comparable to a figure from another, because each has its own dataset, attribute set and counting method. Why Fingerprint Test Tools Disagree About You explains why two honest tools can give you different answers on the same afternoon. For the underlying mathematics of why combinations of ordinary attributes become rare, see Browser Fingerprint Entropy and Anonymity Sets Explained; the arithmetic is not repeated here.
The Time Dimension
Most testers give you a snapshot: one visit, one number. AmIUnique's research framing is different, because a fingerprint that is rare today is only useful for tracking if it stays stable tomorrow. Browser updates, a new font, a different monitor or a changed setting all shift attributes, and that drift is exactly what the FP-STALKER work set out to measure. That is the part Cover Your Tracks does not emphasise. Check the site itself for which history or timeline views it currently offers, since its features may change, but the lesson is stable: uniqueness and persistence are separate questions. A unique fingerprint that changes every week tracks poorly, and a moderately rare one that never changes can track very well.
What It Cannot See
Like every browser-side test, AmIUnique only observes what the browser reports to the page. It cannot tell you which network path your traffic took, and it does not check whether your time zone matches your IP address. For that side of the picture you need a server-side comparison such as BrowserInsight's VPN and proxy check. Broader tool-by-tool context on what these test sites show and omit is in BrowserLeaks Explained.
A Sensible Way to Use It
Run it once with your normal browser, note which attributes the site flags as rare, and treat that as a list of questions: is this setting something I chose, and does it matter to me? Run it again after a browser update and see what moved. Then cross-check with the fingerprint check for per-signal detail. Use the results to understand what is observable, not to chase a "not unique" badge.
Frequently Asked Questions
Is AmIUnique a commercial product?
No. It presents itself as an academic research project on browser fingerprint diversity, and the website's source code is public on GitHub.
Does "unique" mean my browser is fake or tampered with?
No. It means no stored fingerprint matched yours. A completely genuine, unmodified browser with an uncommon combination of settings can be unique.
Can I compare my AmIUnique result with other tools?
Not directly. Each tool has its own dataset, attribute list and counting method, so the numbers describe different populations.
Does AmIUnique check my VPN or IP address?
It reads browser-reported attributes, so it has no view of the network path. Use a server-side VPN and proxy check for that comparison.
