Who is this for?

This Learning Path is for LLM and IoT developers who want to run and interact with AI agents on edge devices like the Raspberry Pi 5. You'll learn how to deploy a lightweight Model Context Protocol (MCP) server and use the OpenAI Agent SDK to create and register tools for intelligent local inference.

What will you learn?

Upon completion of this Learning Path, you will be able to:

  • Deploy a lightweight Model Context Protocol (MCP) server on Raspberry Pi 5 for local AI agent execution.
  • Use the OpenAI Agent SDK to interact with a local AI agent.
  • Design and register custom tools for the agent tasks.
  • Learn about uv - a fast, efficient Python package manager for efficient local deployment.

Prerequisites

Before starting, you will need the following:

  • A Raspberry Pi 5 with a Linux-based OS installed.
  • Familiarity with Python programming and prompt engineering techniques.
  • Basic understanding of Large Language Models (LLMs) and how they are used in local inference.
  • Understanding of AI agents and the OpenAI Agent SDK (or similar frameworks).

Summary

AI-assisted

This summary was drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
You’ll deploy a Model Context Protocol (MCP) server on Raspberry Pi 5 and connect it to an AI agent built with the OpenAI Agent SDK. You’ll install uv and bootstrap a FastMCP server with CPU-temperature and weather tools. Then, you’ll expose the server through an ngrok tunnel and configure an agent on a separate Arm Linux machine. Finally, you’ll connect the server to the agent and prompt to retrieve live values from the Pi and external sources.

Frequently asked questions

AI-assisted

These FAQs were drafted with an approved AI-assisted workflow and reviewed by Arm contributors before publication. Human technical review remains part of the process so the final page reflects engineering rigor, accuracy, and Arm editorial standards.

Close
?
Where do I run the MCP server, and where do I run the AI agent?
Run the MCP server on Raspberry Pi 5. Run the AI agent on your development machine and connect it to the Pi over your network or through the ngrok tunnel.
Do I need `uv` installed on both my Raspberry Pi and my development machine?
Yes. Install uv on Raspberry Pi to set up the MCP server, and install it on your development machine to initialize and run the agent project.
After I run `uv init` for the agent, what files should appear?
uv init creates a .venv virtual environment and a pyproject.toml file in the project directory. Both files indicate that the project environment is ready.
How do I point the agent at the MCP server on my Raspberry Pi?
Add the MCP server address to the agent configuration, using either the ngrok HTTPS forwarding URL or the Pi’s reachable network address. A correct endpoint lets the agent start without connection errors.
How do I know the CPU temperature and weather tools are actually being used?
Prompt the agent to report Raspberry Pi’s CPU temperature or look up weather data. The response includes the requested values. If it doesn’t, verify that the server is running and the connection details match.
Next