Set it up
- 1
Turn on Kubernetes in your AI
In Claude: open Settings → Connectors and add Kubernetes. In ChatGPT: Settings → Apps & Connectors.
- 2
Add it to your app once
Copy the short config below into your app. It just tells your AI where to find Kubernetes.
- 3
Just ask
Tell it what you want in plain words — no special commands.
For example, say
“List all pods in my current namespace and show any that aren't in Running state”
Need the exact config? Show it +
Add to your client
{
"mcpServers": {
"mcp-k8s-go": {
"command": "npx",
"args": ["@strowk/mcp-k8s"]
}
}
}Paste into your client's MCP configuration file.
Get the skill
A ready-made skill that teaches your AI agent to use Kubernetes well. Community Kubernetes specialist skill — Deployments, Helm, RBAC, troubleshooting.
npx skills add jeffallan/claude-skills --skill kubernetes-specialist Quick skills
Copy one, paste it to your AI, watch it work.
Pod Debugging
“The api pod in namespace prod is crash-looping — get the last 100 lines of logs and describe why”
Deployment Scaling
“Scale the checkout deployment to 5 replicas in the production namespace”
Cluster Overview
“List all deployments across namespaces and flag any that have pods not in Ready state”
Good to know
- · The server uses your kubeconfig context — make sure you're not accidentally pointing at production
- · kubectl delete and scale operations are effective immediately — there's no 'are you sure' prompt from the cluster
- · Pods in different namespaces need explicit namespace arguments or queries may look empty
When to use
- ✓ You're debugging a production or staging Kubernetes cluster issue
- ✓ You need to read pod logs and correlate them across services
- ✓ You want to scale deployments or restart pods through conversation
- ✓ You're learning Kubernetes and want AI to explain what's running
When not to use
- × You use Docker Compose, Nomad, or raw containers (use docker-mcp)
- × You want to give AI write access to production clusters without review
- × You need to edit cluster-level RBAC or admission controllers from AI
About
Kubernetes is the system many companies use to run software in production. It groups applications into 'pods,' spreads them across machines, and restarts them if they crash. It's powerful, but it has a steep learning curve — the daily work means typing long kubectl commands and squinting at YAML files. This server lets your AI assistant do that work for you. Ask 'why is the checkout service crashing,' and Claude lists the broken pods, pulls the recent logs, and tells you what looks wrong. Ask 'scale the API up to five replicas because traffic is spiking,' and it makes the change. It can list services, inspect deployments, follow logs, and check the health of anything running in your cluster. For a developer who only occasionally touches Kubernetes, this turns a stressful tool into a conversation. For a team on call, it's a faster way to triage. The trade-off: actions are real and immediate. Use a kubeconfig pointed at a staging or development cluster while you're getting comfortable, and review what the AI is about to run before you approve it for production.
The Kubernetes MCP server provides AI assistants with access to your Kubernetes clusters. List pods, inspect deployments, view logs, manage services, apply manifests, and debug cluster issues — all through natural language.
A powerful tool for DevOps engineers and SREs who want to interact with Kubernetes without memorizing every kubectl command.
Workflows That Use Kubernetes MCP
Works With
Also Consider
FAQ
What does Kubernetes MCP do? +
Kubernetes is the system many companies use to run software in production. It groups applications into 'pods,' spreads them across machines, and restarts them if they crash. It's powerful, but it has a steep learning curve — the daily work means typing long kubectl commands and squinting at YAML files. This server lets your AI assistant do that work for you. Ask 'why is the checkout service crashing,' and Claude lists the broken pods, pulls the recent logs, and tells you what looks wrong. Ask 'scale the API up to five replicas because traffic is spiking,' and it makes the change. It can list services, inspect deployments, follow logs, and check the health of anything running in your cluster. For a developer who only occasionally touches Kubernetes, this turns a stressful tool into a conversation. For a team on call, it's a faster way to triage. The trade-off: actions are real and immediate. Use a kubeconfig pointed at a staging or development cluster while you're getting comfortable, and review what the AI is about to run before you approve it for production.
Do I need to know how to code? +
No. Turn it on in your AI's settings and ask in plain English — no terminal, no coding.
When should I use Kubernetes MCP? +
Reach for it when you need to: You're debugging a production or staging Kubernetes cluster issue; You need to read pod logs and correlate them across services; You want to scale deployments or restart pods through conversation; You're learning Kubernetes and want AI to explain what's running.
When should I avoid Kubernetes MCP? +
Skip it when: You use Docker Compose, Nomad, or raw containers (use docker-mcp); You want to give AI write access to production clusters without review; You need to edit cluster-level RBAC or admission controllers from AI.
How do I install Kubernetes MCP? +
In Claude Code, run: claude mcp add mcp-k8s-go -- npx @strowk/mcp-k8s. Copy-paste configs for Claude Desktop, Cursor, and Windsurf are at the top of this page.
