How to create an MCP server and host it in Docker MCP toolkit?

Creating and Hosting an MCP Server in Docker

Docker provides an excellent way to package and deploy MCP servers — ensuring consistent environments, easy distribution, and isolation.

Step 1: Build an MCP Server

pythonCopy
# mcp_server.py
from mcp.server import Server
from mcp.server.stdio import stdio_server
import mcp.types as types
import asyncio
import os

app = Server("docker-mcp-server")

@app.list_tools()
async def list_tools():
    return [
        types.Tool(
            name="get_env_info",
            description="Get environment information",
            inputSchema={"type": "object", "properties": {}}
        ),
        types.Tool(
            name="echo",
            description="Echo a message back",
            inputSchema={
                "type": "object",
                "properties": {"message": {"type": "string"}},
                "required": ["message"]
            }
        )
    ]

@app.call_tool()
async def call_tool(name: str, arguments: dict):
    if name == "get_env_info":
        return [types.TextContent(type="text",
            text=f"Python version: 3.11, OS: Linux, Container: Docker")]

    if name == "echo":
        return [types.TextContent(type="text",
            text=f"Echo: {arguments['message']}")]

async def main():
    async with stdio_server() as (read_stream, write_stream):
        await app.run(read_stream, write_stream,
                      app.create_initialization_options())

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

Step 2: Create Requirements

textCopy
# requirements.txt
mcp>=1.0.0
anyio>=4.0.0

Step 3: Dockerfile

dockerfileCopy
FROM python:3.11-slim

WORKDIR /app

# Install dependencies
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

# Copy server code
COPY mcp_server.py .

# MCP servers communicate via stdio — no port needed
CMD ["python", "mcp_server.py"]

Step 4: Build Docker Image

bashCopy
docker build -t my-mcp-server:latest .

# Test locally
docker run -i my-mcp-server:latest

Step 5: Docker MCP Toolkit

Docker's official MCP toolkit provides a standardized way to package and run MCP servers:

bashCopy
# Install Docker Desktop (includes MCP support)
# Enable MCP in Docker Desktop settings

# Run an MCP server via Docker MCP toolkit
docker mcp run my-mcp-server:latest

# Docker Desktop automatically discovers MCP servers
# and makes them available to AI clients

Step 6: Configure Claude to Use Docker MCP Server

jsonCopy
{
  "mcpServers": {
    "my-server": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "my-mcp-server:latest"
      ]
    }
  }
}

Docker MCP with Network Access

dockerfileCopy
# For MCP servers that need database or API access
FROM python:3.11-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install -r requirements.txt
COPY mcp_server.py .

# Environment variables for secrets
ENV DATABASE_URL=""
ENV API_KEY=""

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

bashCopy
# Run with environment variables
docker run -i   -e DATABASE_URL="postgresql://user:pass@host:5432/db"   -e API_KEY="your-api-key"   my-mcp-server:latest

Publishing to Docker Hub

bashCopy
# Tag and push your MCP server image
docker tag my-mcp-server:latest yourusername/my-mcp-server:latest
docker push yourusername/my-mcp-server:latest

# Others can use it with:
# docker run -i yourusername/my-mcp-server:latest

Benefits of Docker for MCP