Docker Integration MCP Servers

Docker MCP servers provide interfaces for LLMs to interact with Docker containers and services. These servers enable AI models to manage containers, handle image operations, and assist with Docker infrastructure tasks.

GitHub starsnpm versionnpm downloads

Overview

The MCP Docker Server enables AI models to interact directly with Docker through the Model Context Protocol (MCP). It provides a standardized interface for container operations, allowing AI to assist with container management, image handling, and Docker infrastructure tasks.

Official Server:

Developed and maintained by Model Context Protocol

Key Features

🐳

Container Management

Start, stop, restart, and remove Docker containers.

📦

Image Operations

Build, pull, push, and remove Docker images.

⚙️

Service Orchestration

Manage Docker services and stacks.

💾

Volume Management

Create, inspect, and remove Docker volumes.

Tools

Quick Reference

ToolPurposeCategory
list_containersList all Docker containersContainer Management
create_containerCreate a new Docker containerContainer Management
run_containerCreate and start a Docker containerContainer Management
recreate_containerStop, remove, and run a new containerContainer Management
start_containerStart a Docker containerContainer Management
fetch_container_logsFetch logs from a Docker containerContainer Management
stop_containerStop a Docker containerContainer Management
remove_containerRemove a Docker containerContainer Management
list_imagesList Docker imagesImage Management
pull_imagePull a Docker image from a registryImage Management
push_imagePush a Docker image to a registryImage Management
build_imageBuild a Docker image from a DockerfileImage Management
remove_imageRemove a Docker imageImage Management
list_networksList Docker networksNetwork Management
create_networkCreate a Docker networkNetwork Management
remove_networkRemove a Docker networkNetwork Management
list_volumesList Docker volumesVolume Management
create_volumeCreate a Docker volumeVolume Management
remove_volumeRemove a Docker volumeVolume Management

Detailed Usage

list_containers▶

List all Docker containers.

use_mcp_tool({
  server_name: "docker",
  tool_name: "list_containers",
  arguments: {
    all: true,
    filters: {
      status: ["running"]
    }
  }
});

Returns a list of container objects.

create_container▶

Create a new Docker container without starting it.

use_mcp_tool({
  server_name: "docker",
  tool_name: "create_container",
  arguments: {
    image: "ubuntu:latest",
    name: "my-ubuntu-container",
    command: ["echo", "Hello from Docker"]
  }
});

Returns the ID of the created container.

run_container▶

Create and start a new Docker container (preferred over create_container + start_container).

use_mcp_tool({
  server_name: "docker",
  tool_name: "run_container",
  arguments: {
    image: "nginx:latest",
    name: "my-nginx-server",
    ports: {"80/tcp": "8080"},
    detach: true
  }
});

Returns the ID of the running container.

recreate_container▶

Stop and remove a container, then run a new container.

use_mcp_tool({
  server_name: "docker",
  tool_name: "recreate_container",
  arguments: {
    container_id: "my-old-container",
    image: "my-app:v2",
    name: "my-new-app"
  }
});

Returns the ID of the newly created and running container.

start_container▶

Start a Docker container.

use_mcp_tool({
  server_name: "docker",
  tool_name: "start_container",
  arguments: {
    container_id: "my-stopped-container"
  }
});

Returns success or error.

fetch_container_logs▶

Fetch logs from a Docker container.

use_mcp_tool({
  server_name: "docker",
  tool_name: "fetch_container_logs",
  arguments: {
    container_id: "my-running-container",
    tail: 100
  }
});

Returns the container logs.

stop_container▶

Stop a Docker container.

use_mcp_tool({
  server_name: "docker",
  tool_name: "stop_container",
  arguments: {
    container_id: "my-running-container"
  }
});

Returns success or error.

remove_container▶

Remove a Docker container.

use_mcp_tool({
  server_name: "docker",
  tool_name: "remove_container",
  arguments: {
    container_id: "my-old-container",
    force: true
  }
});

Returns success or error.

list_images▶

List Docker images.

use_mcp_tool({
  server_name: "docker",
  tool_name: "list_images",
  arguments: {
    filters: {
      dangling: false
    }
  }
});

Returns a list of image objects.

pull_image▶

Pull a Docker image from a registry.

use_mcp_tool({
  server_name: "docker",
  tool_name: "pull_image",
  arguments: {
    repository: "ubuntu",
    tag: "22.04"
  }
});

Returns success or error.

push_image▶

Push a Docker image to a registry.

use_mcp_tool({
  server_name: "docker",
  tool_name: "push_image",
  arguments: {
    repository: "myregistry/my-app",
    tag: "latest"
  }
});

Returns success or error.

build_image▶

Build a Docker image from a Dockerfile.

use_mcp_tool({
  server_name: "docker",
  tool_name: "build_image",
  arguments: {
    path: ".",
    dockerfile: "Dockerfile",
    tag: "my-custom-image:latest"
  }
});

Returns success or error.

remove_image▶

Remove a Docker image.

use_mcp_tool({
  server_name: "docker",
  tool_name: "remove_image",
  arguments: {
    image: "my-old-image:latest",
    force: true
  }
});

Returns success or error.

list_networks▶

List Docker networks.

use_mcp_tool({
  server_name: "docker",
  tool_name: "list_networks",
  arguments: {
    filters: {
      driver: ["bridge"]
    }
  }
});

Returns a list of network objects.

create_network▶

Create a Docker network.

use_mcp_tool({
  server_name: "docker",
  tool_name: "create_network",
  arguments: {
    name: "my-custom-network",
    driver: "bridge"
  }
});

Returns the ID of the created network.

remove_network▶

Remove a Docker network.

use_mcp_tool({
  server_name: "docker",
  tool_name: "remove_network",
  arguments: {
    network_id: "my-custom-network"
  }
});

Returns success or error.

list_volumes▶

List Docker volumes.

use_mcp_tool({
  server_name: "docker",
  tool_name: "list_volumes",
  arguments: {}
});

Returns a list of volume objects.

create_volume▶

Create a Docker volume.

use_mcp_tool({
  server_name: "docker",
  tool_name: "create_volume",
  arguments: {
    name: "my-data-volume",
    driver: "local"
  }
});

Returns the name of the created volume.

remove_volume▶

Remove a Docker volume.

use_mcp_tool({
  server_name: "docker",
  tool_name: "remove_volume",
  arguments: {
    volume_name: "my-data-volume",
    force: false
  }
});

Returns success or error.

Installation

{
  "mcpServers": {
    "docker": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-docker"
      ]
    }
  }
}

Common Use Cases

1. Automating Container Deployments

Deploy and manage application containers programmatically:

// Deploy a new web application container
use_mcp_tool({
  server_name: "docker",
  tool_name: "run_container",
  arguments: {
    image: "my-webapp:latest",
    name: "my-webapp-instance",
    ports: {"80/tcp": "8080"},
    detach: true
  }
});

2. Managing Development Environments

Spin up and tear down isolated development environments:

// Start a development database container
use_mcp_tool({
  server_name: "docker",
  tool_name: "run_container",
  arguments: {
    image: "postgres:13",
    name: "dev-db",
    env: ["POSTGRES_DB=mydb", "POSTGRES_USER=user", "POSTGRES_PASSWORD=password"],
    ports: {"5432/tcp": "5432"},
    detach: true
  }
});

3. CI/CD Pipeline Integration

Integrate Docker operations into continuous integration and deployment workflows:

// Build and push a new Docker image as part of a CI pipeline
use_mcp_tool({
  server_name: "docker",
  tool_name: "build_image",
  arguments: {
    path: ".",
    dockerfile: "Dockerfile",
    tag: "my-app:$(BUILD_NUMBER)"
  }
});

use_mcp_tool({
  server_name: "docker",
  tool_name: "push_image",
  arguments: {
    repository: "myregistry/my-app",
    tag: "$(BUILD_NUMBER)"
  }
});

Configuration

The Docker MCP server does not require a connection string. It interacts directly with the Docker daemon available in its environment. Ensure Docker is running and accessible from where the MCP server is launched.

Sources

Related Articles

Elasticsearch MCP Server

Elasticsearch MCP servers enable AI models to interact with Elasticsearch, providing capabilities for searching documents, analyzing indices, and managing clusters.

Linear MCP Server

Linear MCP server enables AI models to interact with Linear's project management system, providing capabilities for issue tracking, project management, team collaboration, and workflow automation.

Integration and Automation MCP Servers

The Integration & Automation category provides integration with automation tools and workflow systems, enabling seamless connectivity and process automation across different platforms and services.