Pure SaaS • Zero Biometric Machines

AI Cloud Face Recognition Attendance System for Modern Workplaces

Eliminate bulky biometric fingerprint terminals and proxy punch-ins. Curiosity HUB transforms standard office laptops, iPads, tablets, and smartphones into high-speed facial recognition attendance kiosks with sub-250ms vector verification, anti-spoof liveness gating, and automatic payroll synchronization.

Sub-250ms Match Latency
99.8% Precision Accuracy
2D FFT Anti-Spoof Liveness
Offline Kiosk Buffering
Bank-Grade 128-d Encryption
How It Works

State-of-the-Art Biometrics Built for the Cloud

Enterprise-grade facial recognition engineered for instantaneous clock-ins across corporate offices, factories, retail chains, and schools.

128-Dimensional Vector AI

Powered by dlib ResNet Euclidean distance matching with in-memory RAM matrix caching. Generates immutable mathematical vector embeddings rather than storing raw photos.

  • Irreversible facial vector arrays
  • Multi-angle face matching (Front/Left/Right)
  • Sub-250ms verification benchmark

Multi-Stage Anti-Spoofing

Prevents buddy punching and fraudulent clock-ins. Our vision pipeline inspects specular reflections, 2D FFT spectral texture frequency, and micro-motion across frames.

  • Rejects photos and printed masks
  • Blocks smartphone screen replays
  • Audit log snapshot with timestamp

Zero-Hardware Kiosk Setup

No specialized biometric readers to purchase or maintain. Run our touchless attendance kiosk on any browser-enabled iPad, tablet, laptop, or mobile device.

  • 5-minute setup with zero technician visits
  • Auto-caching during internet dropouts
  • Central real-time multi-location sync
Technical Specifications

Curiosity HUB Face Recognition Engine Specifications

Industrial-grade performance metrics proven across demanding high-throughput production environments.

Parameter Curiosity HUB SaaS Specification Traditional Hardware Reader (e.g. eSSL/Realtime)
Verification Speed < 250 milliseconds 1.5 – 3.5 seconds (queue bottlenecks)
Initial Hardware Capex ₹0 (Uses existing tablets/webcams) ₹18,000 – ₹35,000 per entrance
Liveness & Anti-Spoofing 2D FFT Spectral Texture + Temporal Micro-motion Basic infrared or single-frame still camera
Biometric Storage 128-d Euclidean Vector Partition (Irreversible) Proprietary device memory or unencrypted templates
Multi-Branch Sync Real-time automatic cloud synchronization Manual USB drive export or complex VPN static IP setup
Leave & Payroll Linkage Native direct sync into 1-click payroll & PDF slips Requires 3rd-party software bridges and manual CSVs
Frequently Asked Questions

Everything You Need to Know About Zero-Hardware Attendance

Clear, transparent answers on biometric accuracy, zero-hardware deployment, offline mode, and payroll automation.

Curiosity HUB is a pure cloud-native SaaS platform. Instead of requiring proprietary fingerprint scanners or expensive dedicated facial terminals (₹15,000–₹35,000 each), our AI recognition engine runs directly on standard hardware you already own—such as office laptop webcams, iPad/Android tablets mounted at entryways, or employee smartphones with GPS geofencing. All processing connects securely to our cloud console without special wiring or maintenance contracts.
Our engine delivers 99.8% verification precision with sub-250ms vector matching. It utilizes high-dimensional Euclidean embeddings (dlib ResNet 128-dimensional vectors) with in-memory RAM caching. Employees simply glance at the camera for less than half a second to verify clock-in or clock-out without bottlenecks at entrances.
No. Curiosity HUB incorporates multi-stage Anti-Spoofing and Liveness Verification. This includes 2D Fast Fourier Transform (FFT) spectral texture frequency analysis to detect printed paper or smartphone screens, specular reflection checking, and multi-frame temporal micro-motion tracking. Still photos, screen replays, and cutouts are instantly rejected.
Yes. Tablet and mobile kiosk modes feature local offline caching. Face verification and timestamp capture execute locally on the device even without active internet. When Wi-Fi or mobile connectivity is restored, all buffered attendance logs automatically synchronize to the central cloud server with tamper-proof timestamps.
We prioritize privacy by design. We do not store raw face photos on public servers. Instead, enrolled facial features are converted into irreversible, 128-dimensional mathematical vector arrays stored in encrypted databases partitioned by tenant ID. It is cryptographically impossible to reverse-engineer a face image from these vector embeddings.
Yes. Administrators can set up unlimited branch offices, warehouses, retail stores, and remote work zones within a single organization account. For roaming or remote employees, our mobile check-in portal enforces GPS geofencing with street-level accuracy, ensuring workers can only clock in within authorized geographic perimeters.
Our integrated Leave Management Software tracks Casual Leave (CL), Sick Leave (SL), Paid Leave (PL), and custom compensatory balances. Employees apply through a passwordless self-service portal, managers approve in 1 click, and public holiday calendars automatically calculate payable days and late-arrival grace periods.
Biometric clock-in timestamps, shift durations, half-day thresholds, and approved leaves flow directly into the Automated Payroll System without manual Excel export/import. The engine automatically computes payable workdays, overtime allowances, unpaid leave deductions, and Indian statutory requirements (PF, ESI, Professional Tax, TDS) with 1-click downloadable PDF payslips.
Any device with a standard camera and modern web browser is supported. This includes Apple iPads, Android tablets (Samsung, Lenovo, etc.), Windows or macOS laptops/desktops with USB webcams, employee Android/iOS phones, and even edge node hardware (like Raspberry Pi 4 with RTSP IP camera streams).
Migration takes under 10 minutes. You can bulk-import employee profiles from your existing HRMS or Excel roster. Employees then receive a self-enrollment link or QR code to upload 3 quick face verification photos (Front, Left, Right) from their smartphone. You can go live immediately without sending technicians or installing software drivers.

Have a specific requirement or architecture question?