Docker

You run a dedicated MCP server to manage Docker with natural language, enabling you to compose containers, inspect running workloads, and handle persistent data with volumes through simple prompts and commands.

How to use

You interact with the MCP server using an MCP client that sends high level intents. Start by describing the containers you want, including a project name and a short description. The system will propose a plan you can review and apply. You can refine the plan, apply changes, or ask for adjustments until you reach the desired state.

Key capabilities include listing, creating, starting, stopping, and removing containers; managing images; creating and removing networks and volumes; and fetching container statistics and logs. Use these actions to build out your docker environment in a guided, conversational way.

How to install

Prerequisites: ensure you have the required runtime installed for the MCP server tools (uv) and a compatible environment for running server processes.

Step 1: Prepare the MCP server configuration for your environment. The following configuration entry configures a local MCP server to run via uv from a repository directory.

"mcpServers": {
  "mcp-server-docker": {
    "command": "uv",
    "args": [
      "--directory",
      "/path/to/repo",
      "run",
      "mcp-server-docker"
    ]
  }
}

Additional setup steps

Optionally, install development tooling or local development helpers to streamline exploring features. If you prefer isolating development from your main workstation, consider using a development container or an environment manager as you progress.

Configuration and security notes

Sensitive data should never be configured directly in containers through the MCP server. Avoid embedding API keys or passwords in prompts or configuration. If you need to pass secrets, use dedicated secret management and mount secrets at runtime.

Troubleshooting and notes

Review created containers carefully before applying changes. Docker can affect the host if used with elevated privileges; the MCP server intentionally disables certain sensitive options to reduce risk.

Development hints

For development work, consider using Devbox to configure your development environment. The recommended approach is to align your setup with local development workflows and test interactions in a safe, isolated environment.

Examples and prompts

Prompts guide you through composing containers or inspecting their state. Use a descriptive project name and a concise description of the desired containers to initiate a plan+apply loop, refining as needed.

Available tools

list_containers

List containers managed by the MCP server, including statuses and IDs.

create_container

Create a new container with specified image, environment, and options.

run_container

Run a container from an image with a user-defined configuration.

recreate_container

Recreate an existing container with updated settings or image.

start_container

Start a stopped container.

fetch_container_logs

Fetch and tail logs from a container.

stop_container

Stop a running container.

remove_container

Remove a container from the environment.

list_images

List available images and their metadata.

pull_image

Pull an image from a registry.

push_image

Push a local image to a registry.

build_image

Build a new image from a Dockerfile or build context.

remove_image

Remove an image from the local store.

list_networks

List networks created for containers.

create_network

Create a new Docker network.

remove_network

Remove a Docker network.

list_volumes

List Docker volumes in use.

create_volume

Create a new Docker volume.

remove_volume

Remove a Docker volume.