Multi-Service MCP Platform

A comprehensive Model Context Protocol (MCP) platform that hosts multiple utility services through a unified HTTP-based API structure. The platform provides various services including transportation information, mindful sentences, weather forecasts, emoji oracle, random timestamps, fortune cookies, and lucky numbers.

Features

Platform Features

• Unified API Structure: Access all services through a consistent URL pattern: http(s)://<domain>/<service>/<version>

• Service Registry: Dynamic service discovery and management

• Service Router: Intelligent request routing to appropriate service handlers

• MCP Protocol Support: Full implementation of MCP protocol methods

• Error Handling: Comprehensive error handling with structured error codes

• Configuration Management: Enable/disable services through configuration

• Health Monitoring: Service status reporting and health checks

• CORS Support: Cross-origin resource sharing for web clients

• Landing Page: User-friendly documentation and service showcase

Available Services

1. MVG Service (/mvg/v1): Munich public transport departures and station information

2. Mindful Service (/mindful/v1): Daily inspirational mindful sentences with rotation

3. Weather Service (/weather/v1): Optimistic weather forecasts with pleasant conditions

4. Emoji Oracle Service (/emoji-oracle/v1): Cryptic guidance through symbolic emoji combinations

5. Timestamp Service (/timestamp/v1): Random timestamp generation within specified ranges

6. Fortune Cookie Service (/fortune/v1): Surreal and cryptic fortune phrases that change hourly

7. Lucky Numbers Service (/lucky-numbers/v1): Mystical number generation for various life scenarios

Related MCP server: Multi-MCPs

Prerequisites

• Python 3.11+

• FastAPI

• Uvicorn

Setup

There are two ways to set up the environment: using venv or uv.

Using venv (standard library)

1. Create a virtual environment:

   python3 -m venv .venv

2. Activate the virtual environment:

   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install the required dependencies:

   pip install -r requirements.txt

Using uv

1. Initialize the project:

   uv init

2. Create a virtual environment:

   uv venv

3. Activate the virtual environment:

   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

4. Install dependencies:

   uv pip install -r requirements.txt

Running the Multi-Service Platform

Start the multi-service platform with:

python start_platform.py

This will display detailed information about available services and endpoints before starting the server.

This starts an HTTP server on http://localhost:8003 with the following endpoints:

• /mcp - Main MCP protocol endpoint

• /services/{service_name}/v{version} - Service-specific MCP endpoints

• /services/{service_name}/v{version}/{method} - Direct service method calls

• /services - List all available services

• /services/{service_name} - Get information about a specific service

• /health - Health check endpoint

• / - Landing page with service information

Service Architecture

The platform uses a plugin-based architecture with the following components:

• ServiceRegistry: Manages service discovery and instantiation

• ServiceRouter: Routes requests to appropriate services

• BaseService: Abstract base class for all services

• Service Implementations: Individual service classes

Adding New Services

To add a new service:

1. Create a new service class that inherits from BaseService

2. Implement the required methods (handle_request, get_tools)

3. Optionally implement prompts support (supports_prompts, get_prompts)

4. Place the service in the services/ directory

5. The service will be auto-discovered by the platform

Example:

from services.base_service import BaseService
from data.models import ToolResult

class MyCustomService(BaseService):
    def __init__(self):
        super().__init__(
            name="custom",
            version="v1",
            description="My custom service"
        )
    
    def get_tools(self):
        return [{
            "name": "my_tool",
            "description": "My custom tool",
            "inputSchema": {
                "type": "object",
                "properties": {},
                "required": []
            }
        }]
    
    async def handle_request(self, method, params):
        if method == "my_tool":
            return self._create_text_response("Custom response")
        else:
            return self._create_error_response(f"Unknown method: {method}")

Service Details

MVG Service

The MVG Service provides real-time information about Munich's public transportation system.

Endpoints

• /mvg/v1

Tools

1. get_departures

   • Description: Get real-time departures from a Munich public transport station

   • Parameters:

     • station (string, required): Name of the station

     • limit (integer, optional): Maximum number of departures to return

     • offset (integer, optional): Time offset in minutes

   • Example:

     {
       "station": "Hauptbahnhof",
       "limit": 5,
       "offset": 10
     }

2. find_station

   • Description: Find stations by name or partial name

   • Parameters:

     • query (string, required): Station name or partial name

     • limit (integer, optional): Maximum number of results to return

   • Example:

     {
       "query": "Marien",
       "limit": 3
     }

Mindful Service

The Mindful Service provides inspirational sentences that rotate throughout the day.

Endpoints

• /mindful/v1

Tools

1. get_mindful_sentence

   • Description: Get an inspirational mindful sentence for the day

   • Parameters: None

   • Example Response:

     "Take a moment to breathe deeply and connect with the present moment."

Features

• Pool of 50+ inspirational sentences

• Daily rotation with 4 different sentences per day

• Consistent sentences within the same day

Weather Service

The Weather Service provides consistently optimistic weather forecasts.

Endpoints

• /weather/v1

Tools

1. get_weather_forecast

   • Description: Get an optimistic weather forecast for tomorrow

   • Parameters: None

   • Example Response:

     "Tomorrow will be a beautiful day with a temperature of 22°C (72°F) and partly cloudy skies. Perfect for outdoor activities!"

Features

• Always predicts pleasant weather conditions

• Temperature ranges between 18-24°C (64-75°F)

• Only sunny or partly cloudy conditions

• No rain, storms, or extreme temperatures

Emoji Oracle Service

The Emoji Oracle Service provides cryptic guidance through emoji combinations.

Endpoints

• /emoji-oracle/v1

Tools

1. get_emoji_oracle

   • Description: Get a cryptic three-emoji combination as guidance

   • Parameters: None

   • Example Response:

     "🌟 🔑 🚪"

Features

• Returns exactly 3 emojis per request

• Draws from a diverse pool of meaningful symbols

• Creates potentially interpretable but cryptic messages

• Different combinations for each request

Timestamp Service

The Timestamp Service generates random timestamps within specified ranges.

Endpoints

• /timestamp/v1

Tools

1. generate_random_timestamp

   • Description: Generate a random timestamp within a specified range

   • Parameters:

     • start_time (string, optional): Start of the time range (ISO 8601 format)

     • end_time (string, optional): End of the time range (ISO 8601 format)

   • Example:

     {
       "start_time": "2025-01-01T00:00:00Z",
       "end_time": "2025-12-31T23:59:59Z"
     }

Features

• Generates random timestamps within specified ranges

• Automatically swaps start/end times if start is after end

• Default range is next day between 10:00 and 16:00

• Returns timestamps in ISO 8601 format

Fortune Cookie Service

The Fortune Cookie Service provides surreal and cryptic phrases that change hourly.

Endpoints

• /fortune/v1

Tools

1. get_fortune_cookie

   • Description: Get a surreal, cryptic fortune phrase

   • Parameters: None

   • Example Response:

     "The silent whisper of forgotten dreams echoes in the corridors of possibility."

Features

• Generates surreal, cryptic phrases

• Same phrase returned within the same hour

• New phrase generated each hour

• LLM integration for content generation

• Supports MCP prompts/list method

Lucky Numbers Service

The Lucky Numbers Service generates "statistically blessed" numbers for various life scenarios.

Endpoints

• /lucky-numbers/v1

Tools

1. generate_blessed_numbers

   • Description: Generates mystically optimized numbers for various life scenarios

   • Parameters:

     • scenario (string, required): The scenario for which you need lucky numbers

       • Options: lottery, decision, restaurant, date, love, career, creativity, health, travel, learning

     • count (integer, optional): Number of lucky numbers to generate (1-10)

     • min_value (integer, optional): Minimum value for generated numbers

     • max_value (integer, optional): Maximum value for generated numbers

     • energy_level (string, optional): Intensity of cosmic energy applied

       • Options: low, medium, high, extreme

   • Example:

     {
       "scenario": "lottery",
       "count": 6,
       "min_value": 1,
       "max_value": 49,
       "energy_level": "high"
     }

Features

• Scenario-specific number generation for 10 different contexts

• Mystical methodology explanations and cosmic alignment calculation

• Customizable parameters (count, range, energy level)

• Formatted responses with emojis and encouraging language

MCP Protocol Support

The platform supports the following MCP methods:

• initialize: Initialize the MCP connection

• tools/list: List available tools across all services

• tools/call: Call a specific tool

• prompts/list: List available prompts (from services that support prompts)

Each service-specific endpoint also supports these methods for that particular service.

API Usage Examples

Using the Main MCP Endpoint

# List all available tools
curl -X POST http://localhost:8003/mcp -H "Content-Type: application/json" -d '{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/list",
  "params": {}
}'

# Call a specific tool
curl -X POST http://localhost:8003/mcp -H "Content-Type: application/json" -d '{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "mindful_get_mindful_sentence",
    "arguments": {}
  }
}'

Using Service-Specific Endpoints

# List tools for a specific service
curl -X POST http://localhost:8003/services/lucky-numbers/v1 -H "Content-Type: application/json" -d '{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/list",
  "params": {}
}'

# Call a specific tool on a service
curl -X POST http://localhost:8003/services/lucky-numbers/v1 -H "Content-Type: application/json" -d '{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "generate_blessed_numbers",
    "arguments": {
      "scenario": "lottery",
      "count": 6
    }
  }
}'

Using Direct Method Calls

# Call a method directly
curl -X POST http://localhost:8003/services/mvg/v1/get_departures -H "Content-Type: application/json" -d '{
  "station": "Hauptbahnhof"
}'

# Get a mindful sentence
curl -X POST http://localhost:8003/services/mindful/v1/get_mindful_sentence

Configuration

The platform can be configured using a JSON configuration file or environment variables.

Configuration File (config.json)

{
  "server": {
    "log_level": "info",
    "port": 8003
  },
  "services": {
    "mvg": {
      "enabled": true
    },
    "mindful": {
      "enabled": true
    },
    "weather": {
      "enabled": true
    },
    "emoji-oracle": {
      "enabled": true
    },
    "timestamp": {
      "enabled": true
    },
    "fortune": {
      "enabled": true
    },
    "lucky-numbers": {
      "enabled": true
    }
  }
}

Environment Variables

• MCP_CONFIG: Path to configuration file (default: config.json)

• MCP_LOG_LEVEL: Logging level (default: info)

• MCP_PORT: Server port (default: 8003)

• MCP_SERVICE_<NAME>_ENABLED: Enable/disable specific services (e.g., MCP_SERVICE_MVG_ENABLED=true)

Claude Desktop Configuration

Add this to your Claude Desktop configuration file (typically ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "multi-service-mcp": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/your/multi-service-mcp-project",
        "python",
        "multi_service_server.py"
      ]
    }
  }
}

Replace /path/to/your/multi-service-mcp-project with the actual path to your project directory.

Testing

Running Tests

# Run all tests
python -m pytest

# Run specific test categories
python -m pytest tests/unit/
python -m pytest tests/integration/

# Run tests for specific services
python -m pytest tests/unit/services/test_lucky_numbers_service.py
python -m pytest tests/integration/test_mindful_service_integration.py

Test Structure

• tests/unit/: Unit tests for individual components

• tests/integration/: Integration tests for service interactions

• tests/performance/: Performance and concurrency tests

MCP Test with Inspector

This command initiates a Model Context Protocol (MCP) test session using the Inspector tool to validate and debug contextual behavior.

npx @modelcontextprotocol/inspector

Docker Support

The project includes a Dockerfile for containerization:

# Build the Docker image
docker build -t multi-service-mcp .

# Run the container
docker run -p 8003:8003 multi-service-mcp

License

This project is licensed under the MIT License - see the LICENSE file for details.