How to beginner · 3 min read

MCP server Docker deployment

Q: MCP server Docker deployment

Deploy an MCP server in Docker by creating a Dockerfile that installs Python and your MCP server code, then run it with docker build and docker run. Use the official mcp Python package and the stdio_server for standard I/O communication inside the container.

Quick answer

Deploy an MCP server in Docker by creating a Dockerfile that installs Python and your MCP server code, then run it with docker build and docker run. Use the official mcp Python package and the stdio_server for standard I/O communication inside the container.

PREREQUISITES

Docker installed on your machine
Python 3.8+
pip install mcp
Basic knowledge of Dockerfile and container commands

Setup

Install Docker from https://docs.docker.com/get-docker/ and ensure Python 3.8+ is available locally for development. Install the mcp Python package with pip install mcp. Prepare your MCP server Python script that uses mcp.server.stdio.stdio_server to run the server over standard input/output.

bash

pip install mcp

Step by step

Create a Python script mcp_server.py that runs the MCP stdio server, then write a Dockerfile to containerize it. Build and run the Docker container to deploy the MCP server.

bash

# mcp_server.py
from mcp.server.stdio import stdio_server

if __name__ == "__main__":
    stdio_server()

---

# Dockerfile
FROM python:3.10-slim

WORKDIR /app

COPY mcp_server.py ./

RUN pip install mcp

CMD ["python", "mcp_server.py"]

---

# Build the Docker image
# Run this in the directory containing Dockerfile and mcp_server.py

docker build -t mcp-server .

# Run the container

docker run --rm -it mcp-server

output

Sending build context to Docker daemon  3.072kB
Step 1/6 : FROM python:3.10-slim
 ---> <image_id>
Step 2/6 : WORKDIR /app
 ---> Using cache
 ---> <image_id>
Step 3/6 : COPY mcp_server.py ./
 ---> Using cache
 ---> <image_id>
Step 4/6 : RUN pip install mcp
 ---> Running in <container_id>
Successfully installed mcp-<version>
Removing intermediate container <container_id>
 ---> <image_id>
Step 5/6 : CMD ["python", "mcp_server.py"]
 ---> Using cache
 ---> <image_id>
Step 6/6 : 
Successfully built <image_id>
Successfully tagged mcp-server:latest

# Running container starts MCP server and waits for stdio communication

Common variations

You can run the MCP server asynchronously by modifying the Python script to use async MCP server APIs if available. To expose the MCP server over TCP inside Docker, add a TCP server wrapper around the stdio server and expose ports in the Dockerfile and docker run command. Use different base images like python:3.11-alpine for smaller containers.

python

# Async MCP server example (if supported)
import asyncio
from mcp.server.stdio import stdio_server

async def main():
    await stdio_server()

if __name__ == "__main__":
    asyncio.run(main())

Troubleshooting

If the container exits immediately, check that mcp_server.py is correctly copied and the CMD is correct. Use docker logs <container_id> to inspect errors. If you see permission errors, ensure Docker has access to your working directory. For networking issues, verify port mappings and firewall rules.

✅

Key Takeaways

Use a minimal Python base image and install the official mcp package inside Docker.
Run the MCP server with stdio_server for simple stdio-based communication inside containers.
Build and run your Docker container with standard Docker CLI commands for easy deployment.
Modify the MCP server script for async or TCP communication as needed for your use case.
Check Docker logs and container status for troubleshooting common deployment issues.

Verified 2026-04

Verify ↗