🐋 Docker MCP server

What is 🐋 Docker MCP server?

It is an MCP server that enables managing Docker containers, images, networks, and volumes using natural language. It integrates with Docker via the Python Docker SDK and is designed for server administrators, tinkerers, and AI enthusiasts.

How to use 🐋 Docker MCP server?

Install by adding the server to your Claude Desktop MCP configuration using either uvx mcp-server-docker or a Docker container that mounts the Docker socket. Then invoke tools and prompts through natural language interactions with the LLM.

Key features of 🐋 Docker MCP server

• Compose containers with natural language prompts
• Introspect and debug running containers
• Manage persistent data with Docker volumes
• List, create, and remove containers, images, networks, and volumes
• Fetch container logs and resource statistics
• Connect to remote Docker engines over SSH

Use cases of 🐋 Docker MCP server

• Deploy a WordPress site with a MySQL backend using natural language
• Manage and inspect containers on a remote Docker host
• Experiment with open-source Docker apps locally via LLM commands

FAQ from 🐋 Docker MCP server

What sensitive data precautions should I take?

Do not configure containers with sensitive data like API keys or passwords, as information exchanged with the LLM is inherently compromised unless the LLM runs locally.

How do I connect to a remote Docker engine?

Set a ssh:// host URL in the DOCKER_HOST environment variable within the MCP server definition.

What Docker options are unsupported?

The server intentionally blocks sensitive Docker options such as --privileged and --cap-add/--cap-drop for safety.

What resources does the server provide?

For each container, the server offers CPU/memory stats and a tail of logs as resources.

How do I install the server via PyPI?

Use uvx mcp-server-docker in your MCP servers configuration after installing uv. Alternatively, build and run the Docker image with the Docker socket mounted.