mctl is a self-service Kubernetes platform that automates infrastructure management via natural language. It integrates GitOps (ArgoCD), secrets management (Vault), and automated service onboarding into
a unified set of 30+ MCP tools. Engineering teams can deploy, scale, and monitor their clusters directly from Claude, Cursor, or Gemini without manual CLI operations.
Overview
Manage Kubernetes with Natural Language
mctl exposes your entire internal developer platform as a set of AI-ready tools. By connecting this MCP server, your AI assistant gains the ability to:
- Deploy & Onboard: Express intent like "Onboard checkout-api from repo X" and let mctl handle the Helm manifests and GitOps commits.
- Observability: Query real-time service health, replica counts, and fetch logs from Loki directly in your chat.
- Database Provisioning: Create PostgreSQL databases on-demand with automated secret injection.
- Scaling & Domains: Manage HPA settings and custom TLS domains without touching kubectl or Cloud Console.
Key Features
- Strict Isolation: Multi-tenancy with team-scoped access via GitHub OAuth.
- GitOps First: Every write operation generates a git commit for full auditability.
- Production Ready: Built on top of industry standards like ArgoCD, HashiCorp Vault, and CloudNativePG.
Quick Start
Visit mctl.ai/mcp to sign in with GitHub and get your access token.
Server Config
{
"mcpServers": {
"mctl": {
"type": "http",
"url": "https://api.mctl.ai/mcp",
"headers": {
"Authorization": "Bearer <YOUR_GITHUB_TOKEN>"
}
}
}
}