School MCP Agent Admin is a privacy-first, local-first management system powered by the Model Context Protocol (MCP). It allows school administrators to query student data and generate PDF reports using natural language, all running on a local LLM. A Local Agentic School Management System using Model Context Protocol (MCP), LangGraph, and Ollama. Feature-rich PoC with SQLite data, PDF reporting, and an AI-driven Streamlit interface.
- Host: Streamlit (Web UI)
- Orchestrator: LangGraph + LangChain
- Client: MCP Python SDK (Subprocess Transport)
- Server: FastMCP (Python)
- LLM: Ollama (Llama 3.1)
- Backend: SQLite
- Install Ollama and run
ollama pull llama3.1. - Python 3.10+.
git clone https://github.com/NxtGenCodeBase/mcp-local-school-orchestrator.git
cd school-mcp-agent-admin
pip install -r requirements.txt
### 🏗️ Project structure
school-mcp-agent-orchestrator/
├── .gitignore
├── README.md
├── requirements.txt
├── database.py # Mock Backend Data Layer
├── server.py # MCP Server (The Tools)
├── orchestrator.py # MCP Client,LangGraph Logic are combined inside app.py file
└── app.py # Streamlit Host Interface
### 🏗️ How it works
- Host (Streamlit) receives your question: "Generate a report for me."
- Host passes this to the Orchestrator (LangGraph/Ollama) to reason.
- The Orchestrator tells the Client (ClientSession): "I need to run the generate_student_report tool."
- The Client sends a JSON-RPC message over stdio to the Server (server.py).
- Server executes the Python code, creates the PDF, and sends the path back to the Client.