JacobStephens2/magisterium_mcp_server

Magisterium API Client

A TypeScript client for interfacing with the Magisterium API.

Environment

Requires at least Node.js 12.20.0 given use of node-fetch (https://www.npmjs.com/package/node-fetch?activeTab=readme)

Setup

1. Install dependencies:

   npm install
   

2. Set up your API key: Create a .env file in the project root and add your API key:

   MAGISTERIUM_API_KEY=your-actual-api-key-here
   

Compiling TypeScript

You have several options to compile the TypeScript code:

Method 1: Using tsconfig.json (Recommended)

npx tsc

This compiles the TypeScript files and outputs them to the dist/ directory.

Method 2: Direct compilation with options

npx tsc --target es2020 --module commonjs magisterium.ts

This creates magisterium.js in the current directory.

Method 3: Simple compilation

npx tsc magisterium.ts

Uses default TypeScript settings and creates magisterium.js in the current directory.

Running the Code

After compilation, you can run the code:

# If compiled with tsconfig.json (Method 1)
node dist/magisterium.js

# If compiled with Method 2 or 3
node magisterium.js

MCP (Model Context Protocol) Tool

This project includes an MCP server that provides a magisterium_query tool for AI assistants to query the Magisterium API directly.

Running the MCP Server

npm run mcp

MCP Tool Usage

The MCP server provides a magisterium_query tool with the following parameters:

• query (required): The question or topic to ask about Catholic Church teaching
• model (optional): The model to use (default: "magisterium-1")
• return_related_questions (optional): Whether to return related questions (default: true)

Example MCP Tool Call

{
  "name": "magisterium_query",
  "arguments": {
    "query": "What does the Catholic Church teach about the Eucharist?",
    "model": "magisterium-1",
    "return_related_questions": true
  }
}

Integrating with MCP Clients

To use this MCP server with compatible AI assistants:

1. Claude Desktop: Add the server configuration to your claude_desktop_config.json:

   {
     "mcpServers": {
       "magisterium": {
         "command": "node",
         "args": ["--loader", "ts-node/esm", "mcp-magisterium.ts"],
         "cwd": "/path/to/magisterium",
          "env": {
            "MAGISTERIUM_API_KEY": ""
          }
       }
     }
   }
   

2. Other MCP Clients: Use the provided mcp-config.json as a reference for configuration.

API Usage

The code makes a direct API call to the Magisterium API:

curl -X POST https://www.magisterium.com/api/v1/chat/completions \
    -H "Authorization: Bearer $MAGISTERIUM_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
    "model": "magisterium-1",
    "messages": [
        {
        "role": "user",
        "content": "What is the Magisterium?"
        }
    ],
    "stream": True
    }'