Welcome to the Building AI Agents with Agno workshop at PyCon Africa 2025! This hands-on workshop will take you through building sophisticated AI agents using the Agno framework, from simple single agents to complex multi-agent teams.

In this workshop, you'll learn how to build AI agents that can:

• Analyze financial data and provide stock insights
• Search and book Airbnb listings using MCP (Model Context Protocol)
• Create intelligent assistants with knowledge bases
• Coordinate teams of specialized agents for complex tasks

You'll also play with AgentOS as a way to serve your agents/teams via a web API.

pycon25/
├── workshop/          # 🎓 Your workspace - complete these exercises
│   ├── a_01_stock_analysis_agent.py
│   ├── a_02_airbnb_mcp_agent.py
│   ├── a_03_agno_assistant_agent.py
│   ├── a_04_news_agency_team.py
│   ├── a_05_code_advice_team.py
│   └── launch_agent_os.py
├── solution/          # 🔍 Reference solutions (if you get stuck!)
│   ├── a_01_stock_analysis_agent.py
│   ├── a_02_airbnb_mcp_agent.py
│   ├── a_03_agno_assistant_agent.py
│   ├── a_04_news_agency_team.py
│   ├── a_05_code_advice_team.py
│   └── launch_agent_os.py
└── README.md

• Python 3.10+
• OpenAI API key (get this from the organizers)
• Agno installed

1. Clone this repository:

   git clone <repository-url>
   cd pycon25

2. Dev Setup: On Unix-like systems:

   ./scripts/dev_setup.sh

   On Windows:

   ./scripts/dev_setup.bat

3. Set up your OpenAI API key:

   export OPENAI_API_KEY="your-api-key-here"

All of the examples in this workshop can be run locally or accessed via AgentOS.

If you want to run the examples via AgentOS, you can do so by running the launch_agent_os.py file.

python launch_agent_os.py

File: workshop/a_01_stock_analysis_agent.py

Build your first AI agent that can analyze stock data using YFinance tools.

What you'll learn:

• Basic agent creation with Agno
• Adding tools and models
• Database integration
• Chat history management

Try asking: "What is the current price of Tesla?"

File: workshop/a_02_airbnb_mcp_agent.py

Create an agent that searches for Airbnb listings using Model Context Protocol (MCP).

What you'll learn:

• MCP integration
• Async agent operations
• Structured thinking and planning
• Real-world API interactions

Try asking: "Find me a 1-bedroom apartment in San Francisco for December 2025"

File: workshop/a_03_agno_assistant_agent.py

Build an intelligent assistant with a knowledge base using vector embeddings.

What you'll learn:

• Knowledge base creation
• Vector database integration (LanceDB)
• Embedding models
• RAG (Retrieval-Augmented Generation)

Try asking: "How do I build agents with knowledge in Agno?"

File: workshop/a_04_news_agency_team.py

Create a team of specialized agents that work together to research and write news articles.

What you'll learn:

• Team coordination
• Agent specialization
• Multi-agent workflows
• Web scraping and content generation

Try asking: "Write an article about latest developments in AI"

File: workshop/a_05_code_advice_team.py

Build a team of code experts that can analyze, review, and improve code.

What you'll learn:

• Advanced team patterns
• Code analysis tools
• Multi-step problem solving
• Agent collaboration strategies

1. Create a new file in the workshop/ directory
2. Create a new agent or team for your idea
3. Test it out by adding the agent/team to the launch_agent_os.py file
4. Share it with everyone!

• Stuck on an exercise? Check the solution/ directory for reference implementations
• Need more examples? Visit docs.agno.com
• Have questions? Ask the workshop facilitators or join the Agno community

• Agno Documentation
• Agno GitHub Repository
• OpenAI API Documentation
• Model Context Protocol (MCP)

Happy coding! 🚀

Built with ❤️ for PyCon Africa 2025