Client Demo

How to Set Up and Use Client Demo

The MCP Server Client Demo is a reference implementation that demonstrates how to build a production-ready, stateless MCP server with streamable HTTP transport using Python and the uv package manager, including an auto tool registry powered by a @mcp_tool decorator, Docker containerization via a Dockerfile.server, and a separate MCP client using the OpenAI SDK. It is designed for developers who want a concrete starting point for building their own MCP server that scales horizontally — rather than exposing domain-specific business tools, it shows the architecture patterns for tool registration, HTTP transport, and environment configuration. Teams adopt it as a scaffolding template when building new MCP servers for internal platform tooling.

git clone https://github.com/S1LV3RJ1NX/mcp-server-client-demo.git
cd mcp-server-client-demo

curl -LsSf https://astral.sh/uv/install.sh | sh
uv sync

cp .env.sample .env
# Edit .env to set your configuration values

uv run python -m server

docker build -f Dockerfile.server -t mcp-server-demo .
docker run --env-file .env -p 8000:8000 mcp-server-demo

{
  "mcpServers": {
    "client-demo": {
      "command": "uv",
      "args": ["run", "python", "-m", "server"],
      "cwd": "/path/to/mcp-server-client-demo"
    }
  }
}

{
  "mcpServers": {
    "client-demo": {
      "command": "uv",
      "args": ["run", "python", "-m", "server"],
      "cwd": "/path/to/mcp-server-client-demo"
    }
  }
}

- "List all the tools available on this MCP server"
- "What tools are registered via the @mcp_tool decorator on this server?"
- "Call the demo tool and show me the response format"
- "Test the server's health endpoint and tell me if it's running correctly"