Control layers selected
Inventory, observability, gateway policy and agent containment can be deployed together or applied at different depths to different use cases.
Enterprise pricing
Kosmoy is scoped around the platform layers, AI workloads, deployment boundary, integrations and support your enterprise requires. The proposal makes each factor explicit before you commit.
How pricing is scoped
The commercial discussion starts with the control boundary, not a generic tier. We separate platform scope, deployment, implementation and external consumption so the total architecture is understandable.
Inventory, observability, gateway policy and agent containment.
The applications, models, agents, teams and environments being governed.
Azure, AWS, GCP, on-premises or isolated Kubernetes environments.
Implementation, integrations, validation, support and operating requirements.
One explicit proposal
The proposal states what Kosmoy controls, where it runs, what implementation is included and which external model or cloud charges remain outside the platform fee.
Inventory, observability, gateway policy and agent containment can be deployed together or applied at different depths to different use cases.
The number and diversity of models, applications, agents, teams, environments and integrations shape the operating footprint.
A standard cloud cluster, multiple regions, on-premises infrastructure or an isolated environment carries different implementation and assurance requirements.
Connectors, migration, validation, operating model, support coverage and service-level requirements are made explicit in the proposal.
Pick the layers
Some AI systems only need to be inventoried. Some need observability. Production traffic may need gateway policy. High-autonomy agents may require a contained runtime and kill switch. One portfolio can use all four depths without forcing every workload into the same package.
Commercial clarity
The commercial proposal states which inventory, observability, gateway, governance and containment capabilities are included.
The target Kubernetes environments and responsibility split for installation, upgrades and operation are documented before contracting.
The proposal identifies the model providers, agent platforms, identity systems, GRC systems and other integrations in scope.
Support coverage, architecture reviews, enablement and any additional assurance or validation work are listed separately.
Where it runs
Kosmoy is Kubernetes-native and single-tenant. The target environment and operating responsibility are part of the commercial scope because deployment architecture affects implementation, assurance and support.
How it works
First contact
Describe the use cases, control gaps and deployment constraints
Discovery
Map the AI estate, stakeholders and required control depth
Technical deep-dive
Review architecture, integrations, security and operating model
Proof of concept
Validate the platform against your own workloads and evidence needs
Commercial proposal
Agree modules, deployment, implementation, support and rollout
Kosmoy is deployed into the customer's Kubernetes and can cover very different control depths: inventory only, observability, runtime gateway policy, or contained agent execution. The proposal therefore states the selected modules, governed workload scope, deployment boundary, implementation and support instead of presenting one number that hides those differences.
The main factors are the control layers selected, the number and diversity of governed applications and agents, the number and type of deployment environments, the integrations required, and the implementation, assurance and support scope.
The Kosmoy proposal covers the platform and the agreed implementation and support scope. Consumption charged by model providers, hyperscalers and customer-owned infrastructure remains visible as a separate external cost rather than being blended into the platform fee.
Yes. Different use cases can sit at different control depths. An enterprise can begin with inventory and observability, add gateway enforcement for production traffic, and use Action Capsules only for agents that require runtime containment. The rollout and commercial scope can follow that sequence.
Kosmoy supports Azure, AWS, GCP, on-premises and air-gapped Kubernetes environments. The proposal states which environments are included and any additional implementation, validation or support requirements associated with the chosen boundary.
Bring the AI workloads, deployment constraints and control gaps you need to cover. We will turn them into a clear platform, implementation and support scope.
Or email sales@kosmoy.com.