VerifyHQ Docs

Service Architecture

VerifyHQ orchestrates multiple internal services to deliver comprehensive identity verification. This page describes how the major components interact.

High-Level Architecture

When you call any VerifyHQ API endpoint, the request flows through the API Gateway, which routes it to the appropriate verification service. Each service is independent but shares common infrastructure for queueing, storage, and event delivery.

Request Flow
Your App
API Client / SDK
API Gateway
Auth, Rate Limit, Routing
Verification Service
Identity / Document / etc.
NTM Engine
AI Models & OCR
Risk Manager
Scoring & Decisions
Event Bus
Kafka / Webhooks
Your Webhook
Callback URL

Core Services

API Gateway

The entry point for all API requests. Handles authentication (API key validation), rate limiting, request logging, and routing to downstream services. Supports idempotency keys for safe retries.

NTM Engine (Neural Trust Model)

The AI backbone of VerifyHQ. NTM handles document OCR extraction, liveness detection scoring, face matching, and document authenticity assessment. It processes images through multiple specialized models and returns structured confidence scores.

NTM Capabilities
Document text extraction (OCR) across 2,000+ African document types, anti-spoofing detection, face similarity scoring, document tamper detection, and age estimation.

Risk Manager

Aggregates signals from all verification services (identity, AML, device, IP) into a unified risk score. Uses configurable rules and ML models to make accept/reject/review decisions. Supports custom risk thresholds per client.

The Risk Manager receives events from each verification step and maintains a real-time risk profile for each user. It powers the Cognitive KYC endpoint which provides holistic risk assessment.

Event Bus

All verification state changes are published to the internal event bus (Kafka-backed). Events are consumed by:

  • Webhook delivery service — delivers signed events to your configured endpoints
  • Risk Manager — updates real-time risk profiles
  • Analytics pipeline — powers dashboard metrics and reporting
  • Watchlist monitoring — triggers re-screening when list updates arrive

Verification Services

Identity Service
Orchestrates multi-step KYC (document + liveness + face match)
Document Service
AI document extraction and authenticity verification via NTM
Liveness Service
Challenge-based photo verification and anti-spoofing
Phone/Email Service
OTP generation, delivery (SMS/email), and verification
Address Service
Proof-of-address document processing and validation
AML Service
Sanctions, PEP, and adverse media screening
Watchlist Service
Continuous monitoring with daily re-screening
Business Service
KYB — registry lookup, directors, UBO screening
Device/IP Service
Fingerprinting, geolocation, and fraud signal detection
Biometric Service
Face encoding, deduplication, and 1:N matching
Mobile Money Service
Account holder name lookup across African providers
Cognitive KYC
Unified AI risk assessment combining all signals

Data Flow & Storage

All verification data is encrypted at rest (AES-256) and in transit (TLS 1.3). Biometric data is stored as irreversible mathematical vectors. PII is isolated in a dedicated encrypted data store with row-level access control.

Document images are retained for the configured retention period (default: 90 days) and then permanently deleted. Verification results and audit logs are retained for 7 years to meet regulatory requirements.

Self-Hosted Deployment

Enterprise customers can deploy the full VerifyHQ stack within their own infrastructure. The self-hosted package includes all core services, NTM models, and the Risk Manager. Contact sales@verifyhq.com for details.