GIM-MCP is an educational interaction management system that implements the Model Context Protocol (MCP) to manage and generate educational content, such as flashcards, with integrated AI capabilities via Ollama. It exposes a RESTful API for chat-based interactions and supports dynamic educational interactions, making it easier to create, organize, and render study materials with AI assistance.

How to use

You interact with GIM-MCP through an MCP-compatible client. Start the server in development or production mode and use the provided REST endpoints to chat with the AI-powered educational resources. You can create and transform flashcards, manage interactions, and render content for study sessions. Use the MCP resources and tools exposed by the server to invoke actions like creating a flashcard, querying for study material, or transforming editor content into a renderer.

How to install

Prerequisites you need before installing the server are clearly defined and must be satisfied.

# Prerequisites
# 1) Node.js (v18 or higher)
# 2) Ollama (for AI functionality)
# 3) Git (to clone the repository)

# Ollama setup (ensure Ollama is running and the required model is downloaded)
ollama pull llama3.1
ollama list

# Clone the project
git clone <repository-url>
cd gim-mcp

# Install dependencies
npm install

# Verify TypeScript configuration
npm run build

Configuration and startup

Environment variables provide runtime configuration. You can customize the API port, Ollama URL, and model by setting these variables in a .env file or exporting them in your shell.

# API server port
PORT=3000

# Ollama URL
OLLAMA_URL=http://localhost:11434

# Ollama model to use
OLLAMA_MODEL=llama3.1

Usage in development mode

Run the server in development mode to enable hot-reload and quick iteration during development.

npm run dev

The server will be available at http://localhost:3000

## Usage in production mode

Build the project for production and then start the compiled server.

npm run build npm start

## Testing the chat endpoint

Test the chat endpoint to send a message and receive a response from the educational content system.

curl -X POST http://localhost:3000/api/chat
-H "Content-Type: application/json"
-d '{"message": "Create a flashcard about mathematics"}'

Or using PowerShell

Invoke-RestMethod -Uri "http://localhost:3000/api/chat" -Method POST -ContentType "application/json" -Body '{"message": "Create a flashcard about mathematics"}'

## Additional sections

This MCP server provides a REST API to interact with educational content and an MCP-compliant interface for resources and tools. It integrates Ollama for local AI models and supports dynamic interactions, including flashcards with text, images, and categories. The project validates inputs with Zod schemas and exposes resources and tools via MCP-compatible endpoints such as interaction://interaction-flashcard and read\_interaction\_flashcard.

## Available tools

### read\_interaction\_flashcard

Invokes the read operation for flashcard interactions, enabling retrieval or rendering of flashcard content via MCP tools.