Myongji University Notice Bot (MCP Server)

smithery badge

An MCP server built with Smithery CLI

The server exposes a tool that can be called by AI agents or language models to fetch and display university announcements directly within a chat interface or other applications.

✨ Features

Real-time Notice Fetching: Uses a Scrapy spider to crawl the official Myongji University notice board (mju.ac.kr) in real-time.
AI Tool Integration: Exposes a get_mju_notices tool that AI agents can naturally call in response to user prompts like "What are the latest announcements from Myongji University?".
Robust and Asynchronous: Built on a modern Python stack, leveraging asyncio, crochet, and subprocess to handle web crawling without blocking the main server loop.
Configurable: The number of notices to retrieve can be specified with a limit parameter.

🛠️ How It Works

An AI agent receives a user request for Myongji University notices.
The agent identifies the get_mju_notices tool as the best way to fulfill the request.
The Smithery MCP server executes the tool.
The tool runs a Python subprocess to invoke a dedicated Scrapy spider.
The Scrapy spider crawls the university's notice page, scrapes the announcements (title, link, date), and outputs the data as JSON.
The tool reads the JSON output, formats it, and returns the list of notices to the AI agent.
The AI agent presents the information to the user in a natural, conversational format.

🚀 Getting Started

Prerequisites

Python 3.10+
uv (or pip) for package management
Git

Installation

Clone the repository:
git clone https://github.com/<your-username>/mju-notice-mcpserver.git cd mju-notice-mcpserver
Create a virtual environment:
uv venv
Install dependencies in editable mode: This command reads the pyproject.toml file and installs all necessary packages, including the local mju_crawler package.
uv pip install -e .