Enterprise facial attendance without storing photos.
High-speed biometric clock-in terminal powered by on-device neural vectors. Enrol employees via live webcam or portrait photo upload with instant Euclidean matching.
Built for enterprise speed & zero compromise privacy
A state-of-the-art attendance system engineered with on-device machine learning.
Sub-Second Recognition
Lightweight 68-point facial landmark aligner and 128-dimensional embedding model running entirely on-device for instant verification in under 300ms.
Active Liveness Detection
Randomised micro-motion challenges, blink detection, and 3D depth-ratio checks permanently block printed photos, screen replays, and silicone masks.
Irreversible Vector Storage
Raw video frames never leave client memory. Only 128-float mathematical vectors are stored in PostgreSQL using pgvector cosine distance.
Workforce Directory
Manage multi-department staff, live camera or picture upload enrollment workflows, job titles, and status controls with role-scoped staff permissions.
Multi-State Event Logging
Clock-in, clock-out, break-start, and break-end events with duplicate suppression windows and real-time confidence scores.
Cryptographic Audit Trails
Every recognized face records match distance, device identifier, liveness index, and microsecond timestamps for compliance audits.
How FaceTime Attendance preserves privacy
Traditional attendance apps upload employee portrait photos to unsecure storage. FaceTime Attendance eliminates this attack vector entirely:
{
"employee_id": "9b1deb4d-3b7d-4bad-9bdd-2b0d7b3dcb6d",
"pose": "front",
"quality": 0.984,
"embedding": [-0.0418, 0.1284, -0.0931, 0.0512, 0.2194, ...128 floats]
}GDPR & CCPA biometric compliant: mathematically one-way.