This is the repository for the LinkedIn Learning course OpenAI Agents SDK. The full course is available from LinkedIn Learning.
AI agents generate plans and perform complex multi-step tasks based on the user input. The OpenAI API provides comprehensive support for building AI agents that use models, tools, knowledge and memory, audio and speech, guardrails, orchestration, and MCP that can be customized to fit specific needs and scenarios. This course explores how to develop advanced agents with the OpenAI API and the Agents Python SDK. Through videos, articles, and hands-on practice in GitHub Codespaces, you’ll define custom agents, set up agent runners, configure handoffs so agents can work together, add guardrails for content moderation, and extend the agents with off-the-shelf OpenAI features, custom functions, and MCP (Model Context Protocol) servers.
See the readme file in the main branch for updated instructions and information.
This repository has three main folders:
./adventurebot/: The folder you'll work with./mcp_server_weather/: An MCP Server you'll incorporate into the project./adventurebot-advanced/: An alternate version of the Adventurebot agent, for reference
To start developing, install all dependencies:
pip install -r requirements.txtYou need an OpenAI API key to run the Agents SDK. Get your key at https://platform.openai.com/api-keys
To use the API key, install the key into your environment:
export OPENAI_API_KEY=<your-key-here>Morten Rand-Hendriksen
Principal Staff Instructor, Speaker, Web Designer, and Software Developer
Check out my other courses on LinkedIn Learning.
