Browser fingerprinting
Identifying a visitor from the combination of subtle browser and device traits, a layer of detection that survives changing your IP.
Fingerprinting works on a principle from information theory rather than from surveillance: no single attribute identifies you, but enough weakly identifying attributes multiplied together do. Your screen resolution narrows the population a little. Your time zone narrows it more. Your exact font list, your graphics card's rendering quirks, your CPU core count, your language preferences, each one divides the remaining crowd again, and after a dozen such divisions the crowd is usually you.
The attributes that carry the most weight are the ones that leak hardware. Draw a shape to a canvas and read the pixels back and the result varies subtly with GPU, driver and font rasterisation. Do the same through WebGL, or process a silent waveform through the audio stack, and you get more of the same. None of these were designed to identify anyone. They identify because hardware and driver combinations are diverse and their imperfections are stable.
For proxy users the decisive fact is that all of this sits ABOVE the network, so changing your address does nothing to it. Rotate through a thousand IPs while presenting one fingerprint and a site correlating the two does not see a thousand visitors. It sees one visitor with a suspicious number of addresses, which is a far stronger signal than simply staying on one IP would have been.
There is a specific trap that catches proxy users more than anyone else, and it is worth naming on its own: the contradiction between what the network says and what the browser says. Your IP geolocates to Frankfurt while your browser reports America/New_York as its time zone and en-US as its language. No real German visitor looks like that. The proxy was fine, the address was trusted, and the browser gave the game away for free.
Which is why fingerprint management is a separate discipline from proxy selection rather than a bonus feature of it. Anti-detect browsers exist to hold a consistent, plausible profile and to keep that profile aligned with the exit address. The right target is not randomness, it is coherence: a profile that is ordinary, internally consistent, and consistent with where the IP says you are.
Frequently asked questions
Does changing my IP change my browser fingerprint?
Not at all, and this is the gap that catches people. Fingerprinting reads the browser and the hardware underneath it, entirely above the network layer. Rotate through a thousand addresses with one fingerprint and a site correlating the two does not see a thousand visitors, it sees one visitor with a suspicious number of addresses.
What is the most identifying part of a browser fingerprint?
The attributes that leak hardware, because hardware imperfections are diverse and stable. Canvas and WebGL rendering vary subtly with GPU, driver and font rasterisation, and audio processing does something similar. None was designed to identify anyone; they identify because the variation is real and consistent.
How do I stop my fingerprint contradicting my proxy?
Align the browser with the exit. If the IP geolocates to Germany, the reported time zone and language should be consistent with that, because no real German visitor reports America/New_York and en-US. Antidetect browsers exist to hold a coherent profile and keep it matched to the address, which is why they are a separate tool from a proxy.
Is randomising my fingerprint a good idea?
Usually not, and it is a common misstep. Randomness is itself unusual, and a fingerprint that changes between page loads describes something no real visitor does. The target is coherence rather than noise: an ordinary, internally consistent profile that also agrees with where your address says you are, held steady for the session.
Back to the full glossary.