natlight/mcp-server-orchestration

MCP Server Orchestration

Demonstration of MCP server orchestration - an intelligent code assistant that manages and coordinates multiple MCP servers using OpenAI Agents SDK.

Overview

Question: Can an MCP server start up and orchestrate other MCP servers?

Answer: YES! This repository demonstrates a sophisticated MCP server (code-assistant-server.py) that orchestrates multiple other MCP servers as internal components to provide intelligent code generation with proper standards compliance and up-to-date library documentation.

Architecture

External MCP Client
        ↓
┌─────────────────────────────────────────┐
│        code-assistant-server.py         │ ← Main MCP Server
│                                         │
│  ┌─────────────────────────────────────┐ │
│  │        OpenAI Agent                 │ │ ← Orchestration Layer
│  │                                     │ │
│  │  ┌─────────────┐ ┌─────────────────┐│ │
│  │  │ Standards   │ │ Context7        ││ │ ← Internal MCP Clients
│  │  │ MCP Client  │ │ MCP Client      ││ │
│  │  └─────────────┘ └─────────────────┘│ │
│  └─────────────────────────────────────┘ │
│                                         │
│  ┌─────────────────────────────────────┐ │
│  │           MCP Server                │ │ ← External Interface
│  │      (External Interface)           │ │
│  └─────────────────────────────────────┘ │
└─────────────────────────────────────────┘
        ↓                    ↓
┌─────────────────┐  ┌─────────────────┐
│ standards-server│  │ context7 server │ ← Internal MCP Servers
│   (subprocess)  │  │   (subprocess)  │   (managed as subprocesses)
└─────────────────┘  └─────────────────┘

Servers

1. Code Assistant MCP Server (code-assistant-server.py)

The main orchestrating server that:

• Starts and manages other MCP servers as subprocesses
• Uses OpenAI Agents SDK for intelligent orchestration
• Communicates with internal MCP servers via JSON-RPC
• Provides intelligent code generation with context from multiple sources

2. Team Coding Standards MCP Server (standards-server.py)

Provides access to team coding standards and best practices as resources.

Key Features

• MCP Server Orchestration: Demonstrates that MCP servers can start and manage other MCP servers
• Intelligent Agent Integration: Uses OpenAI Agents SDK for smart coordination
• Multi-Server Communication: Combines coding standards and library documentation
• Production-Ready: Proper error handling, cleanup, and monitoring
• Multiple Language Support: Python, JavaScript, TypeScript, Java, Go, Rust, SQL, HTML, CSS

Installation & Setup

1. Install dependencies:

pip install -r requirements.txt

2. Set OpenAI API Key:

export OPENAI_API_KEY=your_openai_api_key_here

3. Ensure Docker is running (for Context7 server)

4. Run the Code Assistant Server:

python code-assistant-server.py

Usage Examples

Basic Code Generation

{
  "method": "tools/call",
  "params": {
    "name": "generate_code_with_context",
    "arguments": {
      "request": "Create a REST API endpoint for user authentication",
      "language": "python",
      "libraries": ["fastapi", "pydantic"],
      "include_tests": true,
      "include_docs": true
    }
  }
}

Check Server Status

{
  "method": "tools/call",
  "params": {
    "name": "get_server_status",
    "arguments": {}
  }
}

Testing

Run the test suite:

python test_code_assistant.py

Files

• code-assistant-server.py - Main orchestrating MCP server
• standards-server.py - Team coding standards MCP server
• test_code_assistant.py - Comprehensive test suite
• requirements.txt - Python dependencies
• mcp-config-example.json - MCP client configuration example
• IMPLEMENTATION_SUMMARY.md - Detailed technical documentation

How It Works

1. Startup: Code assistant server starts internal MCP servers as subprocesses
2. Request Processing: External requests trigger the OpenAI Agent
3. Context Gathering: Agent fetches coding standards and library documentation
4. Code Generation: Agent generates context-aware code following best practices
5. Response: Complete solution returned with tests and documentation

Benefits

• Modularity: Each MCP server has specific responsibilities
• Intelligence: Agent makes decisions about context gathering
• Scalability: Easy to add more internal MCP servers
• Reusability: Components can be used independently

License

MIT License - see LICENSE file for details.