Videos
Short demo videos showcasing Apeiro components in action.
Moving Workloads
This demo shows how Platform Mesh and KCP enable policy-driven workload placement across Kubernetes clusters. Workloads are described as SiteWorkload custom resources with declarative placement constraints — such as requiring a specific region or preferring green-energy sites. A Workload Placement Operator evaluates these constraints and schedules workloads to the matching target clusters automatically, without changes to the workload itself.
SimpleCluster Sovereign LLM – Bring Your Own Cluster
This demo illustrates the "two controls, one SaaS service" model: the provider manages the full lifecycle of the LLM service (upgrades, ops, status), while the customer decides where the workload and data actually run by supplying a kubeconfig Secret pointing to their own cluster. The workload and data plane move to the customer-selected host; the SaaS contract and provider-managed lifecycle remain unchanged.
One Chat App. Two Private Models.
This demo walks through a multi-layer architecture built on Platform Mesh: a chat application (Open WebUI) is connected to a private LLM via API bindings as custom resources — LLMInstance, ChatUIInstance, APITokenRequest. The underlying model (e.g. TinyLlama) can be swapped for a stronger one by updating the LLMInstance CR, while the chat application and its API contract stay completely unchanged. The four layers — consumer, provider API, and workload — are each hosted in separate clusters and interact only through the Platform Mesh API abstraction.
