Platform architecture

Engineered as infrastructure. Designed for intelligence.

A layered architecture connects video sources, heterogeneous AI runtimes, event processing, platform services and enterprise applications.

One pipeline. Multiple AI consumers.

Decode a camera stream once wherever practical and distribute shared frames to multiple AI consumers. Runtime selection, worker supervision and distributed inference help the platform grow with your infrastructure.

Shared decodingFrame distributionWorker supervisionHorizontal scalabilityDistributed inference

Shared video processing

A frame manager coordinates camera runtime and frame distribution across multiple models.

Heterogeneous inference

Choose CPU, GPU and supported accelerators according to model compatibility and available infrastructure.

Enterprise operations

Centralize authentication, RBAC, audit logs, settings, system health and integration interfaces.

Technical architecture

Eight layers. One connected platform.

Video Sources

RTSPONVIFUSBVideo filesIP cameras

Video Runtime

Frame ManagerCamera RuntimeWorker Supervision

AI Runtime

Inference SchedulerModel RuntimeHardware Acceleration

AI Models

PersonVehicleFireSmokePPEFaceANPRCustom

Intelligence Layer

EventsSnapshotsRulesAlertsNotifications

Platform Services

AuthenticationRBACUsersSitesAreasSensorsAudit LogsSettings

Observability

Platform healthCamera healthCPU / GPU / RAMStorageWorkersAI runtimeEdge health

Enterprise Interfaces

VIGILIX CoreVIGILIX EyeREST APIWebhooksIntegrations
Heterogeneous inference

Match models to infrastructure.

Choose supported runtimes and execution modes according to hardware availability, model requirements and operating conditions.

VIGILIX CORE / AI DEPLOYMENT

PPE Detection

Warehouse Gate 04

Running

Device: NVIDIA GPU · Execution: Live

Your next layer of intelligence

Bring intelligence into operation.

Explore VIGILIX Core for AI deployment, Eye for outputs and reports, and ORVEXION for infrastructure intelligence.