Device fingerprint
The identity built from a device's hardware and system traits, more stable than a browser fingerprint and harder to change because it reflects the physical machine.
A device fingerprint aims one level deeper than a browser fingerprint: at the machine itself rather than the software presenting it. Screen dimensions and pixel density, CPU core count, available memory, GPU model, sensor availability on mobile, battery characteristics, the exact way the operating system handles time and locale, all of it composes into an identity that persists even when the browser above it changes. Switch browsers on the same laptop and much of the device fingerprint survives.
The distinction from a browser fingerprint is one of durability and is worth holding precisely. Browser-level traits, the user agent, installed fonts, canvas output, belong to the software and can be altered by changing or configuring the browser. Device-level traits reflect physical hardware and are far stickier, which is what makes device fingerprinting valuable to a defender: it identifies the machine that keeps coming back regardless of how the software is dressed up, and it is why creating genuinely separate identities on one physical computer is hard.
For proxy and multi-account work this is the wall behind the wall. Rotating addresses does nothing to it, and even a well-managed browser fingerprint can be undermined if the device traits underneath contradict it, a mobile user agent on a device reporting a desktop's core count and screen, for instance, is a contradiction the address never touches. It is the reason antidetect browsers and, at the strongest, separate physical or virtual machines exist: to make the device layer as consistent and as separable as the browser layer.
The honest limit for a defender is that device signals are noisier and less unique than the marketing suggests: many people own the same phone model with the same specs, so a device fingerprint alone rarely singles out an individual. Its power is in combination, tying a device to behaviour, to accounts, and to a browser fingerprint over time, which is exactly why coherence across all those layers, not perfection in any one, is what keeps identities separate.
Frequently asked questions
What is the difference between a device fingerprint and a browser fingerprint?
A browser fingerprint is built from software traits, the user agent, fonts, canvas output, that change if you change or configure the browser. A device fingerprint targets the physical machine, screen, CPU, GPU, memory, sensors, which persist across browser changes. The device layer is stickier and harder to alter, which is why it is the deeper identity to manage.
Can a device fingerprint identify me across different browsers?
Largely, yes, which is the point of it. Because it reads hardware and system traits rather than the browser, switching browsers on the same machine leaves much of the device fingerprint intact. Combined with behaviour and accounts over time, that lets a site recognise one machine even as the software above it changes, and rotating your address does nothing to it.
How do I create separate identities on one computer?
You have to make the device and browser layers appear genuinely different and internally consistent, which is what antidetect browsers attempt and what separate virtual or physical machines achieve most completely. Changing the browser alone is not enough if the device traits underneath stay identical, since a contradiction between the two is itself the giveaway.
Back to the full glossary.