O'Reilly Live Training - OpenAI AgentKit

Welcome to the O'Reilly Live Training on OpenAI AgentKit! This course will guide you through building AI agents using OpenAI's latest APIs, including structured outputs, file search, and multi-modal capabilities.

Setup

Using uv (Recommended)

This project uses uv, a fast Python package installer and resolver. The Makefile handles most setup automatically.

1. Install uv:

   curl -LsSf https://astral.sh/uv/install.sh | sh

2. One-command setup:

   make all

   This creates a virtual environment in .venv, installs dependencies, and sets up Jupyter kernel.

3. Activate the environment:

   source .venv/bin/activate

4. Setup your OpenAI API key:

   • Get your API key from OpenAI Platform
   • Create a .env file in the project root:

   echo "OPENAI_API_KEY=your-api-key-here" > .env

Using Pip (Traditional Method)

1. Create a Virtual Environment: Navigate to your project directory. Make sure you have Python 3.11+ installed!

   python -m venv .venv

2. Activate the Virtual Environment:

   • On macOS and Linux: source .venv/bin/activate
   • On Windows: .\.venv\Scripts\activate

3. Install Dependencies:

   pip install --upgrade pip
   pip install -r ./requirements/requirements.txt

4. Setup Jupyter Kernel:

   python -m ipykernel install --user --name=openai-agentkit

5. Setup your OpenAI API key: Create a .env file in the project root:

   echo "OPENAI_API_KEY=your-api-key-here" > .env

Remember to deactivate the virtual environment when done: deactivate

Using Conda

• Install anaconda or miniconda
• This repo was tested with Python 3.11
• Create an environment: conda create -n openai-agentkit python=3.11
• Activate your environment: conda activate openai-agentkit
• Install requirements: pip install -r requirements/requirements.txt
• Setup Jupyter kernel: python -m ipykernel install --user --name=openai-agentkit
• Setup your OpenAI API key

Quick Start with Makefile

The project includes a Makefile for common tasks:

# Create virtual environment and install everything
make all

# Clean up environment
make clean

# Add new packages
make add pandas numpy

# Update requirements after manual changes
make env-update

# Show activation command
make activate

Setup your .env file

Create a .env file in the project root:

OPENAI_API_KEY=your-openai-api-key-here

Notebooks

Core Learning Path

The main notebooks are organized in a progressive learning path:

1. Intro to Agents with Responses API - Fundamentals of building agents with OpenAI

2. Simple Chat with Responses API - Building conversational interfaces

3. Structured Outputs for Data Extraction - Extracting structured data with guaranteed schemas

4. Agentic Workflows with Structured Outputs - Building complex agent workflows

5. File Search & RAG with Agentic Retrieval - Implementing retrieval-augmented generation

6. Video Script Generation with Images - Multi-modal agents with vision capabilities

7. Chat with Paper Agent - Building document Q&A agents

8. Research Report Generation - End-to-end research agent system

Additional Resources

• OpenAI API Overview - Comprehensive overview of OpenAI's APIs

• Reference Implementations: Check notebooks/reference-implementations-for-production/ for production-ready examples

Repository Structure

├── notebooks/                              # Main learning notebooks
│   ├── 1.0-intro-agents-responses-api.ipynb
│   ├── 2.0-simple-chat-responses-api.ipynb
│   ├── 3.0-structured-outputs-data-extraction.ipynb
│   ├── 4.0-agentic-workflow-with-struct-out-responses.ipynb
│   ├── 5.0-file-search-rag-agentic-retrieval.ipynb
│   ├── 6.0-video-script-generation-with-images.ipynb
│   ├── 7.0-chat-with-paper-agent.ipynb
│   ├── 8.0-research-report-generation.ipynb
│   ├── openai-api-overview.ipynb
│   └── reference-implementations-for-production/
├── presentation/                           # Course presentation materials
├── assets/                                 # Images, diagrams, and resources
├── demos/                                  # Demo applications
├── requirements/                           # Python dependencies
│   ├── requirements.in                     # Direct dependencies
│   └── requirements.txt                    # Locked dependencies
├── Makefile                                # Automation scripts
└── .venv/                                  # Virtual environment (created by setup)

Key Features

This course covers:

• OpenAI Responses API: Building agentic workflows with the latest OpenAI APIs
• Structured Outputs: Guaranteed JSON schemas for reliable data extraction
• File Search & RAG: Implementing retrieval-augmented generation patterns
• Multi-modal Agents: Working with text, images, and documents
• Production Patterns: Best practices for deploying AI agents

Troubleshooting

Jupyter Kernel Not Found:

python -m ipykernel install --user --name=openai-agentkit

API Key Issues: Make sure your .env file is in the project root and contains:

OPENAI_API_KEY=sk-...

Package Installation Issues: Try upgrading pip first:

pip install --upgrade pip
pip install -r requirements/requirements.txt

Additional Resources

• OpenAI Platform Documentation
• OpenAI Cookbook
• OpenAI API Reference

License

Materials created for O'Reilly Live Training