OpenAIs Python APIs, Agents SDK & Codex App: A Code-Intensive Intro (Video Course)

Description

An example-driven intro to generative AI programming with OpenAIs Python APIs, building single- and multi-agent systems with OpenAIs Agents SDK, and using Codex for vibe coding and AI-assisted software engineering.

Welcome to Paul Deitel's new example-driven course on programming with OpenAIs Python GenAI (generative AI) APIs, building agents using OpenAI's Python and Agents SDKs, and vibe coding and agentic engineering with Codex.

Related Learning:

• Python Fundamentals with Paul Deitel, 2nd Edition video course by Paul Deitel
• Python for Programmers print and e-books by Paul Deitel and Harvey Deitel [Second Edition under development.]

Description:

Large language models (LLMs), generative AI and agents are powering the next generation of AI applications. In this example-driven course led by best-selling author and programming-languages trainer Paul Deitel, youll receive a practical intro to building AI-powered applications using the official OpenAI Python and Agents SDKs, as well as vibe coding and agentic engineering with Codex.

In Part 1, OpenAI Python APIs via the OpenAI Python SDK, Paul demonstrates APIs for text generation, summarization, sentiment analysis, named-entity recognition, natural-language translation and identification, vision capabilities for understanding image content, speech recognition, speech synthesis, image generation and editing, content moderation, and various AI coding tasksgenerating code from a prompt; explaining code; adding type hints, docstrings, and comments to code; generating unit tests; performance tuning; and translating code between languages.

In Part 2, OpenAI Agents SDK, Paul builds single-agent and multi-agent systems that demonstrate the ReAct (Reason + Act) pattern, maintaining context and state for multi-turn conversations, streaming responses, guardrails, using Pydantic for data validation, agent orchestration, extending agents capabilities with hosted and custom tools, and using non-OpenAI models locally via LiteLLM and Ollama. In the agent tools segment, Paul demonstrates:

• Custom function tools and the following hosted tools,
• WebSearchTool for enabling agents to augment models trained knowledge with online contenta form of Retrieval-Augmented Generation (RAG),
• ImageGenerationTool for creating and editing images,
• FileSearchTool for RAG with vector stores,
• CodeInterpreterTool for code generation and execution,
• HostedMCPTool for integrating tools outside the OpenAI ecosystem via Model Context Protocol (MCP),
• ComputerTool for enabling agents to control your local computer, and
• ShellTool for enabling agents to execute shell commands on your local computer.

In Part 3, Vibe Coding and Agentic Engineering with the OpenAI Codex App, Paul demonstrates AI-assisted coding, including code generation, explanation, testing, optimization, documentation, and translating code among programming languages. He describes common coding and software engineering tasks in natural language, lets the Codex agent perform the tasks with human-in-the-loop guidance as appropriate, and discusses the results.

Skill Level:

• Beginner to Intermediate

Learn How To:

• Set up your Python OpenAI programming environment and obtain your OpenAI API key so you can work with the OpenAI APIs, Agents SDK and Codex.
• Leverage your Python skills and the OpenAI APIs to build multimodal AI apps that understand, generate and manipulate text, images, speech, video closed captions and code, and moderate text and images.
• Leverage your Python skills and the OpenAI Agents SDK to build single-agent and multi-agent systems that can autonomously reason, use tools and execute multi-step tasks.
• Use the Codex app (now part of ChatGPT) for vibe coding and agentic engineering tasks.

Course Requirements:

Prerequisites:

• Object-oriented Python programming experience.
• Paul presents this courses content in convenient Jupyter Notebooks using the JupyterLab interface, which you can learn more about at https://jupyterlab.readthedocs.io/en/latest/.
• Get the code and Jupyter Notebooks: https://github.com/pdeitel/OpenAI-APIs-Agents-SDK-and-Codex-Video-Course

Who Should Take This Course:

• Python developers who see exciting generative AI (GenAI) and agentic AI technologies popping up everywhere and want to stay ahead of the curve and enhance career opportunities with a code-intensive intro to them.
• Python developers who want to integrate OpenAI's generative and agentic capabilities into existing Python workflows.
• IT managers contemplating new Python projects that will use GenAI and agentic technologies and who want an example-driven guided survey of OpenAIs Python APIs, Agents SDK and Codex.
• Technical leads or architects evaluating whether and how to incorporate OpenAI APIs and agentic capabilities into their teams products and who want a guided survey of OpenAIs Python APIs, Agents SDK and Codex.

Job titles:

• Python Developer
• Technical Lead
• Architect
• IT Manager

About Pearson Video Training:

Pearson publishes expert-led video tutorials covering a wide selection of technology topics designed to teach you the skills you need to succeed. These professional and personal technology videos feature world-leading author instructors published by your trusted technology brands: Addison-Wesley, Cisco Press, Pearson IT Certification, Prentice Hall, Sams, and Que. Topics include IT Certification, Network Security, Cisco Technology, Programming, Web Development, Mobile Development, and more.  Learn more about Pearson Video Training at https://www.informit.com/video.

Video lessons can be downloaded for offline viewing in the streaming player. Look for the green down-arrow icon in each lesson.

Sample Content

Table of Contents

Part 0: Intro/Setup

Part 1: OpenAI Python APIs

1-1 Text Generation Via the Responses API

1-2 Speech Recognition, Speech Synthesis, and Closed Captions

1-3 Image Generation and Style Transfer

1-4 Content Moderation for Text and Images

1-5 Generating Code with a Codex Model and the Responses API

Part 2: OpenAI Agents SDK

2-1 Single Agent System Python Tutor: Agent, Runner.run(), RunResult

2-2 Conversation State in Agents: previous_response_id, SQLiteSession, conversation_id, result.to_input_list()

2-3 Streaming Text and Events: Runner.run_streamed(), StreamEvent, run_item_stream_event

2-4 Python Tutor with a Model-Backed Input Guardrail: @input_guardrail, GuardrailFunctionOutput, InputGuardrailTripwireTriggered

2-5 Tools

2-5-0 Tools Overview: hosted, local and custom tools

2-5-1 Financial Research Agent: hosted WebSearchTool, custom @function_tool

2-5-2 Image Generation and Editing with ImageGenerationTool

2-5-3 Multi-agent customer service system: FileSearchTool, RAG, vector stores, handoffs between agents

2-5-4 Code Interpreter Tool: CodeInterpreterTool, hosted container

2-5-5 Local MCP SQLite Database: MCPServerStdio

2-5-6 Hosted MCP Weather and Geocoding: MCPServerStreamableHttp

2-5-7 AccuWeather Agent with ComputerTool: AsyncComputer, Playwright for automated web-browser interactions

2-5-8 ShellTool Folder Inspector: Custom executor, human-in-the-loop approval

2-5-9 Local LLM via LiteLLM and Ollama: LitellmModel

2-5-10 Python Code Tutor with Dynamic Instructions: Callable instructions, RunContextWrapper, typed context

Part 3: Vibe Coding and Agentic Engineering with the Codex App

3-0 Overview

3-1 AGENTS.md

3-2 Connecting Codex to Your Project Folder

3-3 Creating a Word-Cloud Code-Generation Script

3-4 Explaining Generated Code

3-5 Adding Type Hints, Docstrings and Comments

3-6 Creating Unit Tests

3-7 Performance Tuning

3-8 Translating Code Between Languages

Part 4: Wrap-Up and Additional References

Updates

Submit Errata