Bot Detection & Threat Intelligence
The Rust Sentry engine scores every event packet in real time using statistical machine learning, impossible travel velocity detection, and live threat intelligence feeds.
“Mechanical bot traffic is automatically flagged by analyzing inter-arrival intervals and hardware concurrency fingerprints.”
Threat Engine Overview
Z-Score Bot Detection
Flags mechanical request rates that deviate more than 3 standard deviations from human baseline activity.
Impossible Travel
Calculates physical speed between consecutive session events using the Haversine great-circle formula.
Threat Database Sync
Synchronizes with Tor exit node directories, VPN lists, and hosting provider ASN registries.
Canvas Fingerprinting
Extracts GPU font-rasterization hashes to identify scrapers spoofing User-Agent strings.
Statistical Z-Score Bot Detection
Human browsing exhibits natural variance in inter-arrival intervals (e.g. 2s, 14s, 5s). Mechanical crawlers fire requests on fixed intervals (e.g. exactly 100ms). The engine flags sessions exceeding ANOMALY_THRESHOLD (default 0.85).
Haversine Impossible Travel
If an event occurs in Dar es Salaam and a subsequent event in the same session occurs in London 5 minutes later, the calculated velocity exceeds commercial flight speeds (> 900 km/h), triggering an impossible_travel security flag.
Hardware Canvas Fingerprinting
The SDK renders an off-screen canvas element during session initialization to capture GPU rendering differences, providing stable device identification without cookies.