Kubernetes MCP Server

Category: Cloud DevOps MCP Servers
Tags: kubernetes, cloud-native, devops, monitoring

Description

Kubernetes MCP Server is a solution that enables LLMs to manage, monitor, and interact with Kubernetes clusters, supporting AI-assisted cloud infrastructure management. It connects to a Kubernetes cluster and allows command execution via CLI or integration with tools like Claude Desktop and mcp-chat.

Features

• Connect to a Kubernetes cluster using the current kubectl context
• List all pods, services, deployments, nodes, and namespaces in the cluster
• Create, delete, and describe pods
• Retrieve logs from pods (supports deployments, jobs, and label selectors)
• Support for Helm v3:
  • Install Helm charts with custom values
  • Uninstall releases
  • Upgrade existing releases
  • Namespace and version specification support
  • Support for custom repositories
• kubectl explain and kubectl api-resources command support
• Fetch Kubernetes events from the cluster
• Port forwarding to pods
• Choose namespace for subsequent commands (command memory)
• CLI integration via mcp-chat or Claude Desktop
• Local development support with hot reload, unit tests, and build scripts
• SSE transport for advanced integrations

Architecture

• Uses a request/response flow between client, transport layer, server, handler, Kubernetes manager, and the Kubernetes API
• Supports both tool operations (like pod management) and resource queries via a modular handler design

Requirements

• kubectl installed and configured with access to a Kubernetes cluster
• Valid kubeconfig file with contexts
• Helm v3 installed (optional if not using Helm features)

License

MIT

Pricing

No pricing information provided. The project is open source and available under the MIT license.

Source

GitHub Repository Package on npm