Kubernetes-MCP
Integrates Kubernetes with Generative AI to manage cloud-native applications, offering automation and insights for Kubernetes management. Facilitates interaction with both local and cloud-based Kubernetes clusters.
Author

lewiesnyder
Quick Info
Actions
Tags
Kubernetes MCP
An MCP for integrating kuberentes with Generative AI
More Information on Model Context Protocol Python SDK and MCP Quickstart
Demo
Prerequisites
- Claude Desktop
- install python ^3.13
- uv
- kubectl
- kubectx and kubens
- docker - for running locally
- kind - for a local k8s cluster
Getting started
If you have a kubernetes cluster:
1. sign in and use kubectx to set the kube context to that cluster.
If you do not have a kuberenets cluster:
1. start docker
2. run infra/startKindCluster.sh to spawn a kind cluster (control plane and two wokers) in docker
Integrate with Claude, Cursor or Windsurf
Tear down
to destroy your kind cluster:
1. run pkill -f "kubectl proxy" to kill the kubectl proxy process.
2. kind delete cluster will delte the kind control plane and worker docker containers
Note: the code here is not meant for production use. User assumes all responsiblity for usage
