Reference Architectures
Curated, end-to-end blueprints that show how multiple building blocks fit together to deliver a complete platform capability. Pick an architecture and get a production-ready platform in minutes. Architectures marked Importable ship their own Terraform code and can be added to meshStack directly.
AI Platform
An opinionated, one-click AI platform: the LiteLLM gateway and a shared ClickHouse deployed into a Kubernetes cluster, with LiteLLM registered as a meshStack platform so a project that lands in an AI landing zone receives a governed model endpoint, a budget and a tracing instance of its own without ordering anything. The cluster comes from the stackit-kubernetes reference architecture; the model backends are whatever the platform team registers behind the gateway.
Azure Kubernetes Platform
A self-service Azure Kubernetes platform that gives application teams their own AKS namespaces with GitHub repositories, GitHub Actions CI/CD pipelines, and a container registry backed by Azure Container Registry — all orchestrated through a single starterkit building block.
STACKIT Kubernetes Platform
A sovereign-cloud Kubernetes platform on STACKIT. One order creates an SKE cluster, gives it an HTTPS ingress with Let's Encrypt certificates and registers it in meshStack as a Kubernetes platform with namespace landing zones. A second order gives application teams self-service namespaces, Forgejo Git repositories, CI/CD via Forgejo Actions and a container registry backed by Harbor.
STACKIT Landing Zone
Bootstraps a self-service-ready STACKIT platform: a dedicated STACKIT resourcemanager folder, a foundation project with a service account, and the STACKIT Project platform with a default landing zone. Three optional layers stack on top of it — hub-and-spoke networking, a Kubernetes platform whose landing zones promise namespaces and a working HTTPS hostname, and an AI platform installed into a cluster ordered through that Kubernetes layer.