Edge-native architecture for low-latency safety monitoring
Dhi intercepts video dynamically at the edge. We deliver sub-200ms intelligence while keeping raw CCTV footage securely on-premises to guarantee privacy-sensitive operational compliance.
Architecture Proof
This page should answer the deployment decision in plain text.
Architecture traffic is high intent. Surface deployment model, latency, privacy, environment fit, and measurable operating outcomes before the reader reaches a demo CTA.
Deployment model
The platform is designed for existing RTSP and ONVIF estates with local compute and native VMS alarm routing.
Latency
This page answers the physical question that matters: whether the alert arrives fast enough to change the outcome.
Privacy posture
Only structured events, approved clips, and system telemetry leave the premises when policy allows it.
Supported environments
The supported-hardware section exists near the fold because architecture buyers need environment fit before they book time.
Measurable outcomes
The performance blocks answer how fast the stack runs, how many streams it supports, and whether it is production-ready.
How Dhi works
A complete pipeline from camera ingest to actionable safety events
Camera ingest
Connect to existing RTSP/ONVIF cameras or integrate with VMS platforms
- RTSP and ONVIF protocol support
- Genetec, Milestone, Avigilon integration
- Multi-stream handling (main + sub-streams)
- Automatic camera discovery and configuration
Edge inference
Real-time pattern detection on industrial edge hardware
- NVIDIA AGX Orin and Jetson NX support
- Industrial x86 edge nodes
- Multi-camera parallel processing
- < 200ms inference latency per frame
Tracking & rules
Multi-object tracking and business rule evaluation
- Cross-camera object tracking
- Zone-based rule engine
- Temporal pattern recognition
- Configurable alert thresholds
Privacy firewall
Policy-based filtering before data leaves the premises
- Raw video stays on-premises
- Structured event generation
- Policy-approved clip extraction
- End-to-end encryption
Event storage
Local and cloud-based event and metadata storage
- Time-series event database
- Searchable metadata
- Configurable retention policies
- Audit trail and compliance logs
Integrations
Connect to incident management and operational systems
- REST API for event streaming
- Webhook notifications
- SIEM and ticketing system integration
- Custom dashboard embedding
Validated on enterprise hardware
Dhi runs on industrial-grade edge nodes designed for 24/7 operation in harsh environments
Ruggedized Infrastructure
The Sentinel-X is purpose-built for the factory floor, the rail yard, and the loading dock. No fans, no moving parts, and an IP67-rated enclosure ensure that your safety intelligence never goes offline—even in the harshest environments.
Built for reliability and scale
Production-grade performance metrics
Continue Exploring
Review privacy and retention controls
See what leaves the premises, what stays local, and how the security review path works.
Use the architecture comparison page
Bring the platform details into a direct edge-vs-cloud evaluation with latency and privacy tradeoffs.
Open the VMS integration guide
Use a concrete proof asset when the next blocker is Milestone, ONVIF, or alarm-routing implementation detail.
Pressure-test the architecture before rollout.
Architecture buyers usually want to review latency, supported environments, privacy boundaries, and integration fit before they book a broader commercial conversation.
We can walk through cameras, compute, VMS integration, and alert routing in one session.
Request a demo
See the flow on a real operating scenario and scope a pilot around one facility or corridor.
See deployment architecture
Review camera ingest, edge inference, alert routing, and what stays on-premises.
Get the implementation checklist
Download the deployment checklist buyers use before green-lighting an industrial AI pilot.
Talk to an engineer
Bring camera count, VMS constraints, latency expectations, and privacy requirements to a technical review.