
Trueface
Enterprise computer vision platform for facial recognition, identity verification, and anti-spoofing. Delivers SDKs and containerized solutions for AI-powered video analytics.
Production-grade AI-powered KYC solution that has successfully onboarded 100K+ users through automated identity verification, face liveness detection, and biometric face matching.
Key Metric
100K+
Users Onboarded
Fully automated — zero manual agent review required.
REF: IDENTITY_VERIFICATION
YAP Pakistan is a production-grade AI-powered KYC platform built for one of Pakistan's fastest-growing fintech super-apps. It processes high-resolution selfies and government-issued identity documents in real time, verifying user identity end-to-end without any manual agent intervention. The system supports CNIC, passport, and driver's licence verification, combined with passive liveness detection to block impersonation and spoofing attacks at every step.
Traditional KYC required human reviewers for each application, creating bottlenecks that slowed onboarding and drove fraud through forged documents and replay-photo attacks. At the scale YAP was targeting—hundreds of thousands of new users—manual review was economically unsustainable and unable to maintain consistent fraud-rejection accuracy.
Liveness & Anti-Spoofing
Multi-frame passive liveness using MediaPipe face mesh and depth-aware texture analysis rejects print attacks, video replay, and 3D masks before document verification begins.
OCR & Biometric Matching
PaddleOCR extracts and validates CNIC fields with post-processing rules; AWS Rekognition cross-references the document photo against the live selfie embedding at sub-second latency.
Real-time liveness detection and anti-spoofing across CNIC, passport, and selfie capture flows.
Biometric face matching with sub-second latency against the extracted document photograph.
Automated fraud-risk scoring and a full compliance audit trail for every verification event.
100K+
Users Onboarded
Fully automated — zero manual agent review required.
<1s
Verification Latency
End-to-end on standard consumer mobile hardware.
0%
False Accepts
In a controlled pilot across 10K high-risk cases.


