Start with onboarding
Choose where it runs
Use the deployment footprint that fits the customer’s policy.
MSSP/MDR · Operate many customers
Run where customer policy allows.
Keep the deployment in your control.
The whole platform, including the AI, runs on-premises, in your cloud, hybrid or fully air-gapped with local models. Agree the required modules and reachable integrations in the deployment design, then validate the customer workflows against that environment before putting the service into operation.
On-premises, cloud, hybrid or fully air-gappedLocal models for air-gapped deploymentsThe same platform in every footprint
Deployment design
Deployment footprint
On-premisesCustomer-operated environment
Your cloudHosting location chosen for the design
HybridSame platform, agreed hosting arrangement
Air-gappedLocal models and reachable internal tools
SelectedAir-gapped deployment
- Customers
- Northstar, Juniper
- Model endpoints
- Local
- Integrations
- Internal network
- Modules
- Agreed in the license
Confirm each required workflow against the environment’s constraints.
Provider build
Search indexClustered search
Virtus AI EngineHorizontal API scaling
Multi-TenantCustomer context across supported work
Capacity and availability requirements belong in the deployment design.
Customer context
NorthstarOwn connector profile
JuniperOwn connector profile
Imperum VaultEncrypted connection credentials
The provider build enables multi-tenant operation.
Providers
Local model endpointRuns inside the environment
Virtus TriageLocal route, where licensed
Virtus PilotLocal route, where licensed
Review the full provider chain for each required workflow.
Air-gapped deployment
- Customers
- Northstar, Juniper
- Model endpoints
- Local
- Integrations
- Internal network
- Modules
- Agreed in the license
Confirm each required workflow against the environment’s constraints.
Workflow review
Required modulesMatch the issued license
ReviewedLocal model endpointsReachable in the environment
ReviewedSupported integrationsConnection tests reviewed
ReviewedIllustrative planning checks, not live diagnostic results.
Air-gapped deployment
- Customers
- Northstar, Juniper
- Model endpoints
- Local
- Integrations
- Internal network
- Modules
- Agreed in the license
Confirm each required workflow against the environment’s constraints.
Agree modules, local model endpoints, installation packages, updates and reachable integrations in the deployment design.
Choose an air-gapped footprint for the illustrated provider deployment.
The provider build uses a search index cluster and a Virtus AI Engine configuration for horizontal API scaling.
Plan local model endpoints and the routes used by licensed AI features.
Check the required modules, model endpoints and reachable integrations against the agreed air-gapped design.
Illustrative customer workflow
01 - What it does
Let customer policy shape the environment.
Choose the hosting footprint before connecting the customer’s workflow. The platform supports on-premises, cloud, hybrid and fully air-gapped environments, with local models available for work that needs to stay inside the deployment.
The AI can run inside the environment.
The full platform and the models behind its AI can install inside your environment. Cloud and hybrid deployments use the same platform with the hosting location changed. For an air-gapped design, agree local model endpoints and reachable integrations alongside the required modules, installation packages and updates. Validate the intended workflows against those constraints before committing to the architecture.
Your team keeps the same platform across supported deployment footprints.
Provider deployment: briefExample
- Footprint
- Air-gapped
- Customers
- Northstar, Juniper
- Model endpoints
- Local
- Integrations
- Internal network
Illustrative customer and configuration. Not customer performance data.
Plan for provider operation
The provider build clusters the search index and configures the AI engine for horizontal API scaling.
02 - How it works in the product
Plan the footprint, then test the workflow.
Hosting location is one part of the design. Model routes and integration reachability determine whether the intended service works within that environment, especially when the deployment cannot reach external services.
- 01Deployment design
Agree the footprint, enabled modules, installation packages and update approach.
- 02Virtus Settings
Configure local model endpoints and review the route for each licensed AI feature.
- 03Marketplace
Configure the supported integrations reachable from the environment and review the reported test results.
- 04Multi-Tenant
Set up customer tenants and profiles after the deployment design is agreed.
For air-gapped environments, agree the required modules, local model endpoints and reachable integrations in the deployment design. Validate each required workflow against those constraints.
03 - How it connects
Place models and tools inside the design.
AI Fabric supplies model configuration and routing. Integrations supplies access to the customer’s tools. Review both when deciding what the service can run in an on-premises, cloud, hybrid or air-gapped environment.
Works with
In the editions
The Portal, Autonomous SOC and SecOps PlatformIncluded
Agreed in the deployment design; enabled modules follow the license.
04 - What it means for your service
Fit the platform to the customer's environment.
Use the same platform across on-premises, cloud, hybrid and air-gapped footprints. Validate model routes and tool access before committing, so the workflows in your proposal fit the agreed environment.
Hosting follows policy
Choose among supported footprints around the customer’s requirements.
Routes get a deliberate review
Check model endpoints and integrations against the intended data boundary.
A service the customer can assess
Use a validated workflow to review the deployment design together.
Start with one customer
- 01Choose one customer workflow
List the tools and licensed AI capabilities it needs.
- 02Map where it runs
Agree hosting, local models and the connections the environment can reach.
- 03Validate the design
Test the workflow against those constraints before committing to the architecture.
Questions about Deployment for providers.
1Where do Imperum’s language models run?
You configure the model providers and their endpoints, including supported cloud services and locally hosted models. Virtus routing can assign different provider chains to features such as Triage and Pilot, with configured fallbacks. Data follows the selected route, so review the full chain when setting residency requirements. Local model hosting alone does not describe every integration’s data flow.
2Does Imperum replace our SIEM or EDR?
Imperum can work alongside an existing SIEM or EDR through configured integrations. Those connections can supply alerts and make supported actions available to investigations and playbooks. Coverage depends on the connector, credentials and configuration, so validate the tools and workflows you want to retain.
3What does a provider deployment look like?
The provider build supports multi-tenant operation with a search index cluster and a Virtus AI Engine configuration for horizontal API scaling. Agree the footprint, licensed modules and reachable integrations in the deployment design; the build does not establish a customer capacity or availability commitment.
4What needs to be agreed for an air-gapped deployment?
Agree the required modules, local model endpoints, installation packages, updates and reachable integrations. Include endpoint connectivity and evidence storage when forensic collection is required, and confirm each workflow against those constraints before committing to the architecture.
Have any other questions? Talk to our team
Your service.
Your deployment design.
Review the footprint, local models and tool access your customer workflows require.