MCP-server

What is MCP-server?

MCP-server is a simple MCP server designed for Moodle manipulation. It connects to Moodle and uses either OpenAI or a local LLM (via ollama) to process user queries.

How to use MCP-server?

Configure your Moodle credentials in the .env file, create a Python virtual environment, install dependencies with pip install -r requirements.txt, then run python client.py server.py for OpenAI or python client_localLLM.py server.py for a local LLM.

Key features of MCP-server

• Shows the due date of assignments.
• Displays unread messages from Moodle.
• Lists uncompleted quizzes.
• Shows all courses you are taking.

Use cases of MCP-server

• Students checking upcoming assignment deadlines.
• Viewing unread messages without logging into Moodle.
• Tracking incomplete quizzes across courses.
• Getting a quick overview of all enrolled courses.

FAQ from MCP-server

What LLM models does MCP-server support?

It supports OpenAI and local LLM models via ollama.

How do I set up MCP-server?

Create a virtual environment, install the requirements from requirements.txt, and fill in your Moodle information in the .env file.

What actions can MCP-server perform?

It can show assignment due dates, unread messages, uncompleted quizzes, and a list of all courses.

Are there separate scripts for different LLMs?

Yes, client.py is for OpenAI and client_localLLM.py is for using a local LLM (ollama).