Enterprise MCP Server Template

This repository is an opinionated, production-oriented architecture template for building MCP (Model Context Protocol) servers using Python. It focuses on maintainability, extensibility, and engineering practices rather than implementing domain-specific capabilities.

This project intentionally contains minimal business logic.

Its purpose is to demonstrate how to structure production-grade MCP servers. Future repositories extend this template with Linux, Docker, Git, Database, and Kubernetes capabilities.

Overview

This template provides an enterprise-ready Model Context Protocol (MCP) foundation. It is designed to act as a platform for exposing infrastructure operations to AI agents and clients via a standardized communication protocol.

Related MCP server: Kroki MCP

Features

• Opinionated & Architecture-first: Built around a strict, layered architecture.

• Capability Registry: A centralized manager for dynamically loading and registering tools, resources, and prompts.

• Strict Decoupling: Business logic is isolated from the server layer and external infrastructure is abstracted via adapters.

• Production-Ready: Pre-configured with structured logging (structlog), health monitoring, dependency injection, and environment configuration.

• Dockerized: Ready for production and local development using Docker and Docker Compose.

• Fast and Modern: Built with Python 3.11+ and relies on uv for blazing fast package management.

Roadmap

This repository establishes the canonical architecture that will natively evolve into:

This repository
      ↓
Linux MCP
      ↓
Docker MCP
      ↓
Git MCP
      ↓
Database MCP
      ↓
Redis MCP
      ↓
Kubernetes MCP

Getting Started

Local Development

# Start development server
make dev

# Run linting
make lint

# Run formatting
make format

# Run tests
make test

Docker

# Build development image
make docker

# Run the dockerized server
docker run -it simple-mcp-server:dev

Testing

This repository uses pytest and uv to manage and run test cases efficiently.

Running Tests

To run the entire test suite, simply use:

make test

Note: Under the hood, this runs uv run pytest.

Test Structure

Tests are organized into logical layers that reflect the enterprise architecture:

• tests/unit/: Tests individual components (e.g., HealthCheck in test_health.py).

• tests/integration/: Tests the interactions between capabilities, the registry, and services.

• tests/e2e/: End-to-end testing of the fully booted server (Transport to Infrastructure).

• tests/fixtures/: Shared test data, mock configurations, and dependency overrides.

Adding New Tests

When contributing new capabilities, place the corresponding test files in the tests/unit/ or tests/integration/ directory using the test_*.py naming convention.