create-containerLABSCO SUMMARY
Container work at the level you actually think about it.
Deploy this stack, show me what is running, tell me why that one is unhealthy — those verbs cover most of a day with Docker, and each is a single call. The logs tool is what turns it from a launcher into a debugging loop.
What it is
A Python MCP server for Docker operations. Four tools cover what a container session usually needs: create one, deploy a compose stack, list what is running, and read a container's logs.
What you get
- A container created and started
- A Docker Compose stack deployed
- Running containers listed with their status
- A container's logs retrieved
Requirements
Docker Desktop or Docker Engine on the machine, Python 3.12 or newer, and the uv package manager. It runs through uvx and acts through the Docker daemon it can reach.
Setup effort
Add the configuration to your client
USE CASES
01
Bringing up a stack you were handed
Deploy the compose stack, then list what came up to confirm every service actually started.
02
Finding out why a container is unhappy
Read its logs and check its status next to the others — the two things you would otherwise open two terminals for.
03
Standing up a single service to try something
Create the container, then read its logs to see whether it came up clean before you build anything on top of it.
04
Checking what is running before you start more
List the containers and their status, which is usually enough to find the port collision before it happens.
4 tools
Listed in project docs. Your client may show a different list.
deploy-composeget-logslist-containersBEFORE YOU INSTALL
What to know first
Tools you get
4 tools named in the project's documentation.
Requires
- uvx on PATH
- Python 3.12+
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
MCP Clients
Claude Desktop — uvx
CLAUDE DESKTOP — UVX · JSON
{
"mcpServers": {
"docker-mcp": {
"command": "uvx",
"args": [
"docker-mcp"
]
}
}
}
