# Docker MCP — Container Management for AI Agents

> MCP server that gives AI agents Docker container management capabilities. Build, run, stop, and inspect containers through tool calls for automated DevOps workflows.

## Install

Merge the JSON below into your `.mcp.json`:

## Quick Use

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

Now ask your AI agent:
```
"List all running containers"
"Start a PostgreSQL container on port 5432"
"Show logs for the api container"
"Build the Dockerfile in ./backend and run it"
```

## What is Docker MCP?

Docker MCP is a Model Context Protocol server that gives AI agents structured access to Docker. Instead of running raw `docker` CLI commands, agents use typed tools for container lifecycle management — list, run, stop, remove, inspect, logs, and build. This enables AI-driven DevOps workflows where agents manage infrastructure autonomously.

**Answer-Ready**: Docker MCP gives AI agents container management via MCP tools. List, run, stop, inspect, build, and read logs through structured tool calls. Enables AI-driven DevOps and automated container workflows. Works with Claude Code and Cursor.

**Best for**: AI agents managing Docker containers for dev/test/deploy. **Works with**: Claude Code, Cursor, Claude Desktop. **Setup time**: Under 1 minute.

## Core Tools

### Container Management

| Tool | What It Does |
|------|-------------|
| list_containers | List running/all containers |
| run_container | Start a new container |
| stop_container | Stop a running container |
| remove_container | Delete a container |
| inspect_container | Get container details |
| container_logs | Read container output logs |

### Image Management

| Tool | What It Does |
|------|-------------|
| list_images | List available images |
| pull_image | Download an image |
| build_image | Build from Dockerfile |
| remove_image | Delete an image |

### System

| Tool | What It Does |
|------|-------------|
| docker_info | System-wide Docker info |
| docker_compose_up | Start compose services |
| docker_compose_down | Stop compose services |

## Use Cases

| Use Case | Example |
|----------|---------|
| Dev Environment | "Spin up PostgreSQL and Redis for local dev" |
| Testing | "Build and run the test suite in a container" |
| Debugging | "Show me the last 100 lines of the API container logs" |
| Cleanup | "Remove all stopped containers and dangling images" |
| Deployment | "Build the Dockerfile and push to registry" |

## FAQ

**Q: Is it safe to give AI agents Docker access?**
A: Use with caution in production. For dev/test, it is very useful. Consider restricting to read-only operations in sensitive environments.

**Q: Does it support Docker Compose?**
A: Yes, includes compose up/down tools for multi-container orchestration.

**Q: Can it build images?**
A: Yes, the build_image tool accepts a Dockerfile path and build context.

## Source & Thanks

> Part of the [Model Context Protocol Servers](https://github.com/modelcontextprotocol/servers) collection.
>
> Docker MCP — Container management for AI agents

<!-- ZH -->

## 快速使用

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

让 AI Agent 管理 Docker 容器。

## 什么是 Docker MCP？

MCP 服务器，AI Agent 通过工具调用管理 Docker 容器生命周期。

**一句话总结**：Docker MCP 服务器，AI Agent 通过工具管理容器（列出/运行/停止/日志/构建），AI 驱动的 DevOps 工作流。

## 核心工具

容器：list/run/stop/remove/inspect/logs
镜像：list/pull/build/remove
Compose：up/down

## 常见问题

**Q: 安全吗？**
A: 开发测试环境非常有用，生产环境建议限制只读。

## 来源与致谢

> [modelcontextprotocol/servers](https://github.com/modelcontextprotocol/servers)

---
Source: https://tokrepo.com/en/workflows/a2214d04-c235-4aad-8d6b-b27888d08547
Author: MCP Hub