Smart fingerprints, explained without the marketing
What a browser fingerprint actually is, why most antidetect tools get it wrong, and the specific tradeoffs we made in Loki.
A browser fingerprint is the set of signals a website can read about your environment without your permission: user-agent, screen dimensions, time zone, installed fonts, canvas-rendering quirks, WebGL hardware strings, audio context, language, platform, and a dozen smaller things. None of these are secret. Together they identify your machine with surprising accuracy.
Antidetect browsers exist to break that identification. They let you present a different fingerprint per profile. The hard part isn't presenting one. It's presenting one that doesn't itself look fake.
The two ways most tools fail
1. Inconsistent surface
A profile claims to be Chrome 124 on macOS 14. Its canvas-rendering signature is from Chrome 121. Its WebGL vendor string says "ANGLE (Apple)" but the renderer string is a Windows GPU. Real browsers don't have these mismatches. Detectors look for exactly that disagreement.
2. Statistical clustering
Random values land in slices of the distribution that don't exist in the wild. If 0.1% of real browsers report 1366×768 and your tool picks resolutions uniformly across a big list, you over-represent that bucket. Detectors group those over-representations and flag them.
What "smart" means in our engine
- Coherent generation: we pick the OS + browser version first, then sample every downstream signal from a real-world joint distribution conditional on that combination.
- Persistence: a profile's fingerprint is stable across sessions. Detection ML hates that — it expects bots to drift.
- Hardware grounding: WebGL and canvas signatures are derived from a small library of real GPUs we've measured, not procedurally generated.
- Honest gaps: where we can't safely spoof a signal (e.g. some WebGPU paths), we surface that and let you decide whether to expose it.
What this means for your workflow
Most operators don't need to think about any of this. The defaults are tuned for the common case (Chrome on macOS / Windows, recent versions, US/EU locales). The fingerprint editor is there for the cases where you need to deliberately match a target environment — for example, a profile that should look like a corporate Windows 11 laptop in a specific time zone.
If you've been burned by an antidetect tool that "worked yesterday and got flagged today," the answer is usually one of the two failures above. We'd rather under-promise on the marketing page and let the engine speak for itself once you're running it.