Skip to content

container-manager-mcp

Manage Docker, Docker Swarm, Podman, and Kubernetes through a typed MCP server and an integrated A2A agent — local or across a fleet of hosts over SSH, or across multiple Kubernetes clusters via kubeconfig contexts.

Official documentation

This site is the canonical reference for container-manager-mcp, maintained alongside every release.

PyPI MCP Server License GitHub

Overview

container-manager-mcp exposes the Docker, Podman, and Kubernetes engines as a compact set of action-routed MCP tools an agent or IDE can call, plus a Pydantic-AI graph agent for conversational orchestration. It provides:

  • DockerManager / PodmanManager / KubernetesManager — typed Python facades over the Docker, Podman, and Kubernetes (official kubernetes client) engines covering images, containers, volumes, networks, Swarm services, Compose stacks, and the full Kubernetes workload/config/networking/storage/RBAC/cluster/governance/ observability surface.
  • Action-routed MCP tools — consolidated, togglable tool modules (info, image, container, volume, network, swarm, system, compose, plus Docker Swarm, Podman, and 8 themed Kubernetes tool modules) that minimize LLM context overhead. See Kubernetes for the full k8s tool breakdown.
  • Multi-context parallel operation — cm_multi_context fans a single call out across several Docker, Podman, Swarm, and/or Kubernetes contexts at once.
  • Zero-script multi-host control — route any Docker/Podman operation to a remote host over a standard SSH channel, with no Docker TCP socket exposed, driven from a unified inventory.yml (.yaml legacy fallback); remote Kubernetes clusters are reached instead via kubeconfig contexts (K8S_CONTEXTS) — see Multi-Host.
  • Ontology-driven KG ingestion — cm_ingest_inventory maps live Docker/Swarm/ Kubernetes inventory into typed OWL/RDF nodes in the epistemic-graph Knowledge Graph.

Explore the documentation

  • Installation — pip, source, extras, and the prebuilt Docker image.
  • Deployment — run the MCP and agent servers, Docker Compose, Caddy + Technitium.
  • Usage — the MCP tools, the DockerManager Python API, and the CLI.
  • Overview — ecosystem role, enterprise readiness, and architecture.
  • Multi-Host — zero-script Docker-over-SSH control plane, and the Kubernetes kubeconfig-context equivalent.
  • Kubernetes — the 8 cm_k8s_* tools, patch_resource, true core Services, and multi-context.
  • Concepts — the CONCEPT:CMGR-* registry.

Quick start

pip install "container-manager-mcp[all]"
container-manager-mcp              # stdio MCP server (default transport)

Run it as a network server:

container-manager-mcp --transport streamable-http --host 0.0.0.0 --port 8000

See Installation and Deployment for the full matrix (PyPI extras, Docker image, all transports, the agent server, reverse proxy, and DNS).