Silent blind spots
A camera drops off the network and nobody notices until footage is needed for an investigation. Recording gaps are discovered weeks late, when they are no longer fixable.
PURPOSE-BUILT FOR VISION AI
Nebula OS — the Vision AI operating system for your camera estate
Nebula OS is the operating system underneath OrionAlerts. It boots on a single appliance or server and runs video management, AI video analytics, camera health monitoring, asset records, and AMC workflows as one connected system instead of five disconnected tools.
One stack, not five tools
VMS, AI analytics, health, assets and AMC ship as modules of the same OS, on the same data.
Boots like an appliance
Power on and the video, AI, storage and health services come up together — no manual stitching.
Works with the cameras you own
A universal device layer speaks ONVIF, RTSP and OEM SDKs, so the estate stays brand-neutral.
Core Engines
Video EngineRecording EngineAI RuntimeHealth EngineAPI EngineDevice Manager
Cloud Layer
Orion CloudCommand CenterAlertsAnalyticsRemote Management
Cameras → SDKs → Drivers → Nebula OS → Cloud / Command Center
Core Engines
Video EngineRecording EngineAI RuntimeHealth EngineAPI EngineDevice Manager
Cloud Layer
Orion CloudCommand CenterAlertsAnalyticsRemote Management
Cameras → SDKs → Drivers → Nebula OS → Cloud / Command Center
Overview
What is Nebula OS?
Most surveillance deployments are assembled, not designed. A VMS from one vendor records the video. An analytics box from another vendor watches a subset of the cameras. A spreadsheet holds the asset list. A WhatsApp group holds the maintenance history. Nothing shares a database, so nobody can answer a simple question — which cameras are down, who owns them, and are they still under warranty — without opening four systems.
Nebula OS collapses that stack. It is a Linux-based operating system built specifically for Vision AI infrastructure: the device layer, video engine, recording engine, AI runtime, health engine and API engine are all part of the same managed environment, released and updated together. The capability modules — video management, AI analytics, health monitoring, asset tracking and AMC — are surfaces on top of one shared record of every camera, site and event.
The practical result is that a detection, a camera outage, an asset record and a maintenance ticket all refer to the same device object. An AI alert can tell you the camera it came from is out of warranty. A health failure can raise a service request against the vendor who installed it. That correlation is not an integration project — it is how the operating system is built.
THE PROBLEM
The problem with a bolted-together surveillance stack
Fragmented tooling is the reason large camera estates quietly stop working. These are the failure patterns Nebula OS is designed to remove.
The integration tax
Every new capability means another server, another licence, another integration to maintain. Upgrade one component and the others need re-testing.
No single source of truth
Asset lists live in spreadsheets, warranties live in email, and service history lives in people's memory. Audits and vendor disputes become archaeology.
Analytics that nobody trusts
AI bolted onto a subset of cameras produces alerts with no operational context, so operators learn to ignore them and the investment stops paying back.
THE MODULES
What runs inside Nebula OS
Five capability modules, one operating system, one shared record of every camera, site and event.
Video management
Video management (VMS) built into the OS
Live monitoring, recording, playback and multi-site control run as a native service of Nebula OS rather than as a separate video management system you have to integrate.
Explore the VMS module- Live wall and playback across sites from one operator console
- Recording and retention policies managed per camera, zone or site
- Searchable incident timelines and evidence export for investigations
- Role-based access so operators, managers and auditors see the right scope
AI analytics
AI video analytics on the same video pipeline
The AI runtime reads decoded frames directly from the video engine, so analytics run on the cameras you already record — not on a parallel stream into a separate analytics appliance.
Explore AI video analytics- Real-time detection for intrusion, restricted areas, PPE and safety events
- Zone and schedule rules so alerts match how the site actually operates
- Events written to the same record as the camera, site and asset they came from
- GPU-accelerated inference sized to the channel count of the node
Health monitoring
Camera health monitoring as a system service
A health engine continuously checks cameras, recording, storage, services and connectivity. Because it sits inside the OS, it sees failures the network alone cannot — a camera that pings but has stopped recording, for example.
Explore health monitoring- Device, stream and recording status tracked continuously per camera
- Storage capacity and retention risk surfaced before footage is lost
- Server, service and connectivity checks with a built-in watchdog
- Uptime and availability reporting for the whole estate
Asset tracking
Asset records that match the live estate
Every device the OS talks to is also an asset record — location, ownership, serial number, lifecycle stage and documents — so the inventory cannot drift away from reality the way a spreadsheet does.
Explore asset tracking- Single register for cameras, recorders, servers and network devices
- Site, floor, zone and department mapping for every deployed device
- Ownership and custody history for transfers, handovers and audits
- Warranty documents, invoices and manuals attached to the device itself
AMC & maintenance
AMC, warranty and maintenance workflows in the same place
Because health failures and asset records already live in the OS, maintenance stops being a separate discipline: a failing camera can raise a service request against the right vendor, under the right contract, automatically.
Explore AMC management- AMC and warranty coverage tracked against each protected device
- Preventive maintenance schedules across sites and asset classes
- Service requests from intake through vendor dispatch to closure
- SLA tracking, renewal alerts and complete per-asset service history
Built into the OS. Not bolted on later.
Nebula OS combines device connectivity, camera SDKs, video processing, GPU acceleration, APIs, recording, AI runtimes, system health, and cloud connectivity into one managed environment.
- Linux Core
- Universal Drivers
- OEM SDKs
- ONVIF
- RTSP
- GPU Acceleration
- API Engine
- Recording Engine
- Watchdog
- Secure OTA
- Cloud Sync
Boot Like an Appliance
Power on and Nebula OS automatically initializes the video and AI stack.
Multi-Brand Ready
Connect supported camera ecosystems through a unified device layer.
GPU Optimized
Designed for hardware-accelerated video decoding and AI inference.
Self-Monitoring
Watchdog, device health, storage health, and service monitoring built in.
Controlled Updates
Managed OS and service updates with rollback-ready deployment architecture.
Cloud Connected
Synchronize health, alerts, configuration, and operations with Orion Cloud.
DEPLOYMENT
Deployment models
Nebula OS runs where the video already is. The same operating system image and the same modules are used in every model — only the placement of the node changes.
Single plants, campuses and facilities
On-site node
Nebula OS runs on an on-premise appliance or server next to the cameras. Video and AI inference stay inside the site network; only health, alerts and configuration metadata sync outward.
- Video never leaves the site
- Works through constrained or intermittent WAN links
- Lowest bandwidth footprint
Multi-site estates and regional operations
Distributed nodes + Orion Cloud
Each site runs its own Nebula OS node for video, recording and inference, while Orion Cloud provides the central command view — estate-wide health, alerts, assets and AMC across every location.
- Per-site autonomy, central oversight
- One command view across regions
- Remote configuration and managed updates
System integrators and AMC providers
Partner-managed
Integrators deploy and operate Nebula OS nodes on behalf of their customers, using the health, asset and AMC modules to run service delivery across multiple client estates.
- Estate visibility per client
- Service and SLA workflows built in
- Evidence for warranty and contract claims
COMPARISON
Nebula OS vs a traditional surveillance stack
The difference is not a longer feature list. It is that these capabilities share one system, one device model and one release cycle.
| Dimension | Traditional VMS + NVR + add-ons | Nebula OS |
|---|---|---|
| Architecture | Separate VMS, analytics box, monitoring tool and spreadsheets, joined by integrations | One operating system with modules on a shared device and event model |
| AI analytics | A parallel stream into a separate appliance, usually on a subset of cameras | An AI runtime reading the same decoded frames the recorder already uses |
| Camera health | Network ping at best; recording failures found during investigations | Device, stream, recording and storage health checked continuously by the OS |
| Asset & warranty records | Spreadsheets and email threads that drift away from the live estate | Every connected device is an asset record with ownership, warranty and history |
| Maintenance | Raised manually, after someone notices, against whichever vendor is remembered | Driven by health events against the contract already attached to the device |
| Updates | Component-by-component, with re-testing of every integration | Managed OS and service updates with a rollback-ready deployment path |
| Camera choice | Constrained by whichever ecosystem the VMS favours | Brand-neutral through a universal device layer (ONVIF, RTSP, OEM SDKs) |
Architecture
Traditional VMS + NVR + add-ons
Separate VMS, analytics box, monitoring tool and spreadsheets, joined by integrations
Nebula OS
One operating system with modules on a shared device and event model
AI analytics
Traditional VMS + NVR + add-ons
A parallel stream into a separate appliance, usually on a subset of cameras
Nebula OS
An AI runtime reading the same decoded frames the recorder already uses
Camera health
Traditional VMS + NVR + add-ons
Network ping at best; recording failures found during investigations
Nebula OS
Device, stream, recording and storage health checked continuously by the OS
Asset & warranty records
Traditional VMS + NVR + add-ons
Spreadsheets and email threads that drift away from the live estate
Nebula OS
Every connected device is an asset record with ownership, warranty and history
Maintenance
Traditional VMS + NVR + add-ons
Raised manually, after someone notices, against whichever vendor is remembered
Nebula OS
Driven by health events against the contract already attached to the device
Updates
Traditional VMS + NVR + add-ons
Component-by-component, with re-testing of every integration
Nebula OS
Managed OS and service updates with a rollback-ready deployment path
Camera choice
Traditional VMS + NVR + add-ons
Constrained by whichever ecosystem the VMS favours
Nebula OS
Brand-neutral through a universal device layer (ONVIF, RTSP, OEM SDKs)
WHO IT IS FOR
Who Nebula OS is built for
Security & control-room leaders
Know that every camera you are accountable for is actually recording, and investigate incidents without moving between systems.
Plant & operations teams
Use the cameras already installed for safety and process visibility, with alerts that carry the site context operators need to act.
IT & infrastructure
Run one managed platform with one update path instead of maintaining integrations between four vendors' products.
System integrators & AMC partners
Deliver and service multi-client estates with health evidence, asset records and SLA tracking built into the platform you deploy.
FAQ
Nebula OS — frequently asked questions
What is Nebula OS?
Nebula OS is the Vision AI operating system that powers OrionAlerts. It is a Linux-based platform that runs video management, AI video analytics, camera health monitoring, asset tracking and AMC workflows as modules of one system, on one shared record of every camera, site and event.
How is Nebula OS different from a VMS?
A VMS is an application for viewing and recording video. Nebula OS is the operating system underneath it. Video management is one module; the AI runtime, health engine, device manager, asset register and maintenance workflows are part of the same environment, released and updated together rather than integrated after the fact.
Do I need to replace my existing cameras?
No. A universal device layer connects supported IP cameras and recorders through ONVIF, RTSP and vendor SDKs, so existing camera investments can be brought onto the platform rather than replaced.
Does video have to leave my site?
No. In the standard deployment, video and AI inference stay on the on-site Nebula OS node. Orion Cloud synchronises health status, alerts, configuration and operational metadata for central management, without requiring raw video to be uploaded.
Does Nebula OS need a GPU?
Nebula OS is designed for hardware-accelerated video decoding and AI inference, so GPU acceleration is used where AI analytics are enabled. The right node sizing depends on channel count and which detections are running — our team sizes this per deployment.
Can Nebula OS manage multiple sites?
Yes. Each site can run its own node for video, recording and inference, while Orion Cloud provides a single command view of health, alerts, assets and maintenance across the entire estate.
How are updates handled?
The OS and its services are updated together through a managed, rollback-ready deployment path, including secure over-the-air updates for connected nodes — so a platform upgrade is a single controlled change rather than a component-by-component migration.
Who is Nebula OS built for?
Organisations running camera estates where surveillance is operationally critical — manufacturing, logistics and warehousing, energy and utilities, campuses, healthcare, retail and public infrastructure — along with the system integrators and AMC partners who deploy and service those estates.
See Nebula OS running on your cameras
We will walk through the architecture, size a node for your channel count, and show the modules working on a live estate.
Prefer to read first? Browse the solution modules