What is MCTL?
MCTL is an AI-native infrastructure management platform that lets you control Kubernetes infrastructure through natural language, REST APIs, or a developer portal.
The Problem
Managing Kubernetes is complex. Teams need to:
- Write and maintain Helm charts, Kustomize overlays, and ArgoCD configs
- Handle networking, TLS, DNS, and ingress rules
- Set up RBAC, resource quotas, and namespace isolation
- Monitor deployments, troubleshoot failures, and perform rollbacks
This requires deep platform engineering expertise that most teams don't have.
The Solution
MCTL abstracts the complexity behind a unified control plane:
- Natural language — connect Claude, Cursor, or VS Code via MCP and manage infrastructure conversationally
- REST API — integrate with CI/CD pipelines, scripts, and custom tooling
- Developer Portal — Backstage-powered UI for browsing services, tenants, and operations
- GitOps — every change is a Git commit, synced by ArgoCD with full audit trail
How It Works
- You make a request — via MCP tool, API call, or portal action
- MCTL API processes it — validates auth, checks RBAC, generates the desired state
- GitOps repo is updated — the change is committed to
mctl-gitops - ArgoCD syncs — detects the diff and applies it to the cluster
- Operation completes — you can track progress via the operation ID
Every action is an async operation with a trackable ID. You can check status, view logs, and see the full history of changes.
Core Principles
GitOps-first — the Git repository is the single source of truth. No imperative kubectl apply commands.
Multi-tenant — each team gets isolated namespaces with RBAC policies and resource quotas. No cross-tenant access.
AI-native — built from the ground up for AI interaction. The MCP server exposes 39 tools covering the full platform lifecycle.
Self-healing — the mctl-agent watches AlertManager and automatically creates PRs to fix common issues.