What Is Browser Fingerprinting?
How websites can recognize a browser without cookies, an account, or anything saved on your device.
The core idea
Every browser answers technical questions as part of displaying web pages: which fonts can it use, which graphics processor renders images, what resolution is the screen, how does the audio engine process signals. Each answer is mundane on its own. But combine enough of them and the resulting profile — a browser fingerprint — can distinguish one browser installation from most others. Unlike a cookie, a fingerprint is not saved on your device; it is recomputed from your browser's observable behavior every time, so clearing cookies or using private browsing does not reset it.
Why fingerprinting exists
Fingerprinting has both defensive and tracking uses. Banks and login systems use device recognition to spot suspicious sign-ins — if your account is suddenly accessed from a browser that looks nothing like yours, that is a useful fraud signal. Advertising and analytics companies have used the same techniques to recognize visitors across sites after browsers restricted third-party cookies. The technique is identical; the purpose differs, which is why browser vendors treat fingerprinting surface area as a privacy problem while being unable to remove the underlying APIs, since real features depend on them.
The main signal families
| Signal | What it reads | Learn more |
|---|---|---|
| Canvas | Pixel output of a fixed drawing, shaped by hardware, drivers, and font rendering | Canvas fingerprinting |
| WebGL | Graphics vendor and renderer strings, extensions, and limits | WebGL fingerprinting |
| Audio | Exact numeric output of the audio engine for a fixed signal | Audio fingerprinting |
| Screen | Resolution, pixel ratio, color depth, viewport | Screen fingerprinting |
| Fonts | Which fonts are installed, inferred from rendered text measurements | Font fingerprinting |
Real fingerprinting scripts also read user-agent details, language and time-zone settings, hardware concurrency, touch support, and more. This site's analyzer focuses on the five families above because they are measurable, explainable, and demonstrate the principle clearly.
Why combinations matter more than single values
A screen resolution of 1920×1080 describes an enormous number of devices. A particular GPU renderer string describes fewer. A specific set of 30 installed fonts, fewer still. Fingerprinting works by intersection: each signal cuts the set of matching browsers down, and the combination of many moderately-informative signals can be far more distinguishing than any one of them. This is also why partial defenses can disappoint — blocking one signal while leaving five others intact may not change the overall picture much.
What a fingerprint is not
A fingerprint identifies a browser installation, not a person by name. It does not reveal your identity, address, or browsing history by itself. It also is not guaranteed to be unique: many corporate machines with identical hardware, settings, and fonts can produce identical fingerprints. And it is not permanent — browser updates, driver updates, and new fonts change fingerprints over time, which is a genuine practical limit on tracking accuracy.
See your own signals
The best way to understand fingerprinting is to look at your own browser's answers. Run the analyzer, read the explanations, then try the before-and-after comparison to see which defenses actually change which signals. When you are ready to act, the reduction guide covers realistic options.