Noctua Labs NOCTUA LABS

Changelog

v2.4.1 — Engine
  • Added JA4+ transport fingerprint spoofing layer with automatic TLS ClientHello reconstruction
  • Expanded persona database to 2,847 hardware profiles (added M3 Ultra, RTX 5090, Intel Arc B580)
  • Fixed VTC phase drift on sessions exceeding 6 hours
  • Improved canvas noise injection precision from ±3px to ±0.4px
v2.4.0 — Engine
  • Introduced `session_export` and `session_import` commands for portable .nsvault archives
  • Added `network_intercept` command: full HTTP/HTTPS request/response interceptor
  • Launched 120-command Python SDK — 15 new commands added this release
  • MONOLITH-class evasion benchmark: 99.97% oracle pass rate across 847 test vectors
v1.8.0 — Axiom Zero
  • Deployed hardware physicality layer to 300+ edge nodes globally
  • Added MONOLITH-class engine detection module (DTC phase-lock analysis)
  • Reduced P99 edge latency from 18ms to 12.4ms
  • Added real-time threat map fed by live session telemetry
v2.3.2 — Engine
  • Added `tor_rotate` command with automatic exit node validation
  • Improved Hawkes Process kinematics model: WPM accuracy ±1.2 vs real human baseline
  • Fixed WebGL readback timing on AMD GPU profiles
v1.7.5 — Axiom Zero
  • Launched behavioral session scoring: per-session risk score 0.0–1.0
  • Added WebRTC leak detection and IP consistency validation
  • New honeypot pipeline: Tier 3 actor intent logging to encrypted Blobs store
v2.3.0 — Engine
  • Initial public release of Noctua Engine binary
  • 105-command SDK at launch
  • Hardware-accurate persona generation for 1,200+ device profiles
  • Encrypted session architecture with zero disk footprint
Subscribe to release notes →