Platform · Deployment

Where your data sits is your call.

Public bodies cannot send resident data to a jurisdiction they do not control. Regulated firms often cannot either. We treat that as the starting constraint rather than an upgrade you pay extra for, so every capability is available in a deployment that never leaves your perimeter.

Four models. The same product in each.

Sovereign cloud. Hosted in the region you choose and run by us. The fastest way to a production agent, with data residency in the region you choose and a documented processing register.

Private VPC. Inside your own cloud tenancy. Your account, your network boundary, your security controls, our software.

Your own data centre. On your hardware, under your operational control.

Disconnected network. No external connection at all. Open models run on your own GPUs, so no inference call ever leaves the building. This is the configuration for organisations that cannot allow any external connection.

Nothing is held back for the cloud

The distinction between these models is where the software runs, not what it can do. There is no governance feature, connector class or evaluation tool that exists only in our hosted environment. An installation on a disconnected network is the full product.

Environments and lifecycle

Development, staging and production are separate in every deployment model. Every save is a version with a readable diff, and publishing to production requires approval.

Getting there

How long it takes to reach production depends on the process, how many systems it touches and how quickly connections can be approved on your side. We give an estimate after mapping and fix it in the scope document.

Roles, approval gates, environments and retention are configured before anything reaches production. Compliance is designed in, not retrofitted after the pilot succeeds.