Skip to main content

Helm MCP Server

Python 3.12+ MCP License: Apache 2.0 Helm GitHub Discord

An MCP server that gives AI assistants full control over Helm — from chart discovery to production deployments. Any MCP-compatible assistant (Claude, Cline, or your own agent) can search charts, validate values, dry-run before deploying, monitor rollouts, and roll back when things go sideways.

What it does​

Discovery & search — Search charts across Bitnami, ArtifactHub, and custom repos. Pull metadata, versions, READMEs, and values schemas through MCP resources.

Installation & lifecycle — Install, upgrade, rollback, and uninstall releases. Dry-run first to see what would change. Use custom values, multiple values files, and extra CLI args.

Validation & safety — Validate values against JSON schemas before deploy. Render and validate manifests ahead of time. Check dependencies and cluster prerequisites. Get installation plans with resource estimates.

Monitoring & status — Watch deployment health after install or upgrade. Get release status and revision history. List releases across namespaces in one call.

Multi-cluster — List and switch Kubernetes contexts from kubeconfig. Use namespaces for isolation.

Built-in guidance — Workflow guides, security checklists, and troubleshooting runbooks ship as MCP prompts. Upgrades, rollbacks, and incident response are covered.

Architecture​

Quick start​

Prerequisites​

RequirementDescription
DockerFor containerized deployment (recommended)
Python 3.12+For local installation
Helm CLIInstallation Guide
kubectlInstallation Guide
Kubernetes AccessValid kubeconfig with cluster access

Docker Quick Start​

docker run --rm -it \
  -p 8765:8765 \
  -v ~/.kube/config:/app/.kube/config:ro \
  talkopsai/helm-mcp-server:latest

That's it. The server listens on http://localhost:8765/mcp.

Connect your MCP client​

{
  "mcpServers": {
    "helm-mcp-server": {
      "url": "http://localhost:8765/mcp",
      "description": "Helm MCP Server for managing Kubernetes workloads via Helm"
    }
  }
}

Security​

Don't hardcode secrets in values files — use Kubernetes Secrets or external secret managers. Use namespaces for isolation.

Set MCP_ALLOW_WRITE=false when you want the assistant to explore, validate, and plan — but not actually change anything in the cluster. Dry-runs are always allowed regardless.

Project structure​

helm-mcp-server/
├── helm_mcp_server/           # Main package
│   ├── tools/                 # MCP Tools
│   │   ├── discovery/         # Chart search and info
│   │   ├── installation/      # Install, upgrade, rollback, uninstall
│   │   ├── validation/        # Values and manifest validation
│   │   ├── kubernetes/        # Cluster operations
│   │   └── monitoring/        # Deployment monitoring
│   ├── resources/             # MCP Resources
│   ├── prompts/               # MCP Prompts
│   ├── services/              # Business logic
│   ├── server/                # FastMCP server setup
│   ├── static/                # Static documentation
│   └── config.py              # Configuration management
├── tests/                     # Test suite
└── pyproject.toml             # Project configuration

Next steps​