Set it up
- 1
Turn on Docker in your AI
In Claude: open Settings → Connectors and add Docker. In ChatGPT: Settings → Apps & Connectors.
- 2
Add it to your app once
Copy the short config below into your app. It just tells your AI where to find Docker.
- 3
Just ask
Tell it what you want in plain words — no special commands.
For example, say
“List all my running Docker containers with their status and exposed ports”
Need the exact config? Show it +
Add to your client
{
"mcpServers": {
"MCP_DOCKER": {
"command": "docker",
"args": ["mcp", "gateway", "run"]
}
}
}Paste into your client's MCP configuration file.
Quick skills
Copy one, paste it to your AI, watch it work.
Local Environment Cleanup
“List all stopped containers, dangling images, and unused volumes — then remove them”
Container Debugging
“Show me the last 200 lines of logs for the 'api' container and tell me if there are any errors”
Quick Container Ops
“Restart the postgres container and tail its logs until it's ready to accept connections”
In the wild
“I spent the evening testing servers for DuckDuckGo, Grafana, and GitHub. Watching an LLM hit those tools directly from plain text is honestly wild”
“having a dedicated build service that CC can control through MCP vs running Docker itself because it can then restart the container and test”
Good to know
- · Requires Docker Desktop 4.59+ with the MCP Toolkit feature enabled — the `docker mcp` CLI plugin ships with recent Docker Desktop
- · The gateway needs access to the Docker socket — this grants root-equivalent privileges on the host
- · Running against a remote Docker host requires TLS setup or SSH tunneling
- · Destructive actions like 'prune' remove data permanently — always confirm before running
When to use
- ✓ You're managing local development containers and want faster control
- ✓ You're debugging a containerized app and want AI to read the logs
- ✓ You're cleaning up stale images, volumes, or containers after a project
- ✓ You want to inspect container state without memorizing docker commands
When not to use
- × You use Podman, containerd, or Kubernetes directly (use kubernetes-mcp for k8s)
- × You need to give AI permission to run arbitrary containers on production hosts
- × You're building multi-stage production deployments — use dedicated CI/CD tooling
About
Imagine telling your AI: 'Stop all running containers and clean up dangling images' or 'Show me the logs for the redis container' — and it manages Docker for you. This server connects your AI to Docker so it can list, start, stop, and inspect containers, images, volumes, and networks through conversation.
The Docker MCP server provides AI assistants with control over Docker environments. List and manage containers, pull and inspect images, manage volumes and networks, view logs, and execute commands inside containers — all through conversation.
Highly useful for DevOps workflows, local development, and infrastructure management with AI assistance.
Workflows That Use Docker MCP
Status, logs, restart, scale — through chat. No more `docker ps | grep | awk` reflexes.
GitHub, Git, Docker, Kubernetes and Sentry in one chat — review, ship, and debug without tab-switching.
Works With
Also Consider
FAQ
What does Docker MCP do? +
Imagine telling your AI: 'Stop all running containers and clean up dangling images' or 'Show me the logs for the redis container' — and it manages Docker for you. This server connects your AI to Docker so it can list, start, stop, and inspect containers, images, volumes, and networks through conversation.
Do I need to know how to code? +
No. Turn it on in your AI's settings and ask in plain English — no terminal, no coding.
When should I use Docker MCP? +
Reach for it when you need to: You're managing local development containers and want faster control; You're debugging a containerized app and want AI to read the logs; You're cleaning up stale images, volumes, or containers after a project; You want to inspect container state without memorizing docker commands.
When should I avoid Docker MCP? +
Skip it when: You use Podman, containerd, or Kubernetes directly (use kubernetes-mcp for k8s); You need to give AI permission to run arbitrary containers on production hosts; You're building multi-stage production deployments — use dedicated CI/CD tooling.
How do I install Docker MCP? +
In Claude Code, run: claude mcp add MCP_DOCKER -- docker mcp gateway run. Copy-paste configs for Claude Desktop, Cursor, and Windsurf are at the top of this page.
