Browser Identity &
Bot Detection Glossary

A technique used to identify and track individual web browsers by gathering technical parameters exposed during web visits. These include Canvas 2D/WebGL rendering entropy, installed fonts, WebAudio FFT waveforms, Screen geometry, and CPU core counts. Unlike cookies, browser fingerprints persist across private browsing modes and storage clearings.

The process of measuring physical hardware artifacts that only genuine consumer silicon can produce. Axiom Zero measures sub-microsecond CPU cache eviction timing, physical display VSync phase jitter, and GPU subpixel rasterization noise to confirm the presence of real hardware. Virtual machines, headless cloud instances, and software emulation layers fail these physical measurements.

An advanced standard for profiling incoming TLS client connections based on ClientHello parameters. It inspects TLS version, cipher suite order, extensions, ALPN values, and TCP window size to produce a deterministic hash. Automated scripts using Python requests or standard Node.js HTTPS modules exhibit distinct TLS signatures that do not match genuine web browsers.

An extension to OAuth 2.0 (RFC 9449) that cryptographically binds access tokens to a specific asymmetric private key held by the client browser. Axiom Zero leverages hardware-backed DPoP keypairs stored in device TPMs to ensure that session tokens stolen via XSS or network intercept cannot be replayed from authorized devices.

A continuous 0–100 risk score computed at the network edge by Axiom Zero during incoming connection evaluation. A score below 30 represents verified organic human traffic; scores between 30 and 65 trigger transparent tarpits/throttling; scores above 65 activate deception responses (GHOST) or immediate edge blocks.

The discipline of detecting automated browser instances (such as Puppeteer, Playwright, Selenium, or modified Chromium/Firefox builds) operating without a physical display interface. Detection targets include missing navigator plugins, anomalous WebGL extension lists, flatline VSync timing, and lack of neuromotor mouse acceleration.

Mathematical modeling of human neuromotor input patterns. Genuine human mouse trajectories follow Hawkes Point Process kinematics with stochastic excitation and Gaussian inter-keystroke timing distributions. Synthetic mouse movements (Bezier curves, linear interpolations) are immediately flagged by biometrics evaluation.