Build AI Agents with Model Context Protocol (MCP) by Pearson

Course details

Explore ways you can construct AI agents using Model Context Protocol (MCP) by leveraging the similarities it has with HTTP. Learn about the creation of MCP and why it's essential for overcoming traditional walled garden strategies. Explore the architecture of MCP servers and the range of components and capabilities they offer. Gain hands-on experience building MCP servers in Python, Java, and Node.js, as well as testing these servers with MCP Inspector, Postman, and Clawed Desktop. Discover ways MCP enables AI agents to interact naturally with databases to provide real-time information and insights. Plus, integrate MCP with SQL databases like PostgreSQL to solve common enterprise problems and showcase the potential of AI in optimizing workflows. Ideal for developers who are eager to enhance their AI development skills, this course equips you to build efficient AI agents with the robust MCP standard and transform how your services and datasets interact.

Concepts

Introduction

Build AI agents with Model Context Protocol (MCP) - Introduction

Get Started with Model Context Protocol (MCP) for AI Agents

Learning objectives
Learn the domains of information
Lab - Use AI to get real-time information - The weather
Learn why MCP was created
Understand why walled garden strategies fail
Remember lessons learned from HTTP
Learn how MCP is similar to HTTP
Get familiar with MCP client applications
Lab - Use AI with MCP to get real-time information - The weather

Learn the Architecture of MCP Servers

Learning objectives
Learn the basic terms and terminology
Get familiar with the message protocol
Understand the types of transports used - STDIO and Streamable HTTP
Recap - MCP client applications
Understand when to apply tools vs. resources
Learn the purpose of prompts in your MCP servers
Learn more about transports

Deep Dive - Create an MCP Server in Python

Learning objectives
Focus on the implementation details
Learn the dependencies
Compare UV vs. PIP
Create the MCP server
Set the transport
Declare your tool
Declare your prompts
Lab - Run the Python MCP server in the MCP Inspector

Deep Dive - Create an MCP Server in Java

Learning objectives
Focus on the implementation details
Learn the dependencies
Create the MCP server
Set the transport
Declare your tool
Declare your prompts
Lab - Run the Java MCP server in Postman

Deep Dive - Create an MCP Server in Node.js

Learning objectives
Focus on the implementation details
Learn the dependencies
Create the MCP server
Set the transport
Declare your tool
Declare your prompts
Lab - Run the Node.js MCP server in the Claude Desktop

Use MCP with the PostgreSQL Database

Learn the solution
Configure your MCP server
Lab - Run the PostgreSQL MCP server in the Claude Desktop
Learning objectives
Discuss the problem

Summary

Build AI Agents with Model Context Protocol (MCP) - Summary