OpenAI Codex is an AI coding assistant that helps you write, edit, and understand code using natural language. In this course, you’ll learn how to use Codex as a practical tool in your daily development work—specifically through the command-line interface (CLI). While Codex is also available as a web-based assistant in ChatGPT and as an extension for editors like VS Code, we’ll focus on the CLI. This is the most direct way to use Codex for real projects, and all hands-on practice in this course will use the CLI.

At the heart of Codex is GPT-5-Codex, a specialized version of OpenAI’s latest language model. Unlike general-purpose models, GPT-5-Codex is trained not just on code, but on real engineering tasks and workflows. This allows it to understand complex instructions in plain English, generate high-quality code in many programming languages, and even refactor, debug, or explain code with full awareness of your project’s context. Because it’s grounded in real-world engineering problems, GPT-5-Codex is especially good at understanding your intent and producing practical, working solutions.

Codex supports a wide range of programming languages, including Python, JavaScript, TypeScript, Go, Java, and more. It can work with different file types and is designed to be flexible enough for most modern software projects.

Codex CLI brings the capabilities of GPT-5-Codex directly to your terminal. Here are some of the things you can accomplish:

• Generate new code: Describe a function or feature, and Codex will write it for you.
• Refactor or update code: Ask Codex to improve, optimize, or modernize existing code.
• Write and update tests: Quickly generate unit tests for your modules.
• Fix bugs: Point out a problem, and Codex can suggest and apply a fix.
• Explain code: Unsure what a piece of code does? Codex can walk you through it.
• Automate repetitive tasks: Batch rename variables, update documentation, or perform other routine changes.

Codex is especially useful for automating tedious or repetitive work, speeding up prototyping, and helping you get unstuck when you’re not sure how to approach a coding problem. For example, you might ask Codex to “add input validation to the process_data function in utils.py” or “convert this script to use async/await.” Codex will analyze your code, suggest the necessary changes, and show you a preview. You decide whether to accept, modify, or reject the suggestion.

Codex CLI is designed to work with your local files and projects. When you give it an instruction, it uses the context of your codebase to tailor its suggestions. For best results, run Codex in the root directory of your project, so it can access all relevant files and dependencies. Codex reads your code, understands the structure of your project, and proposes changes that fit naturally with your existing codebase.

You interact with Codex by typing instructions in your terminal, and it responds with proposed changes that you can review and approve. This means you stay in control of your code at all times, and you can use Codex as a true assistant in your day-to-day development.

To use Codex CLI on your own machine, you would typically follow these steps:

1. Install Codex CLI:

   If you encounter permission errors, it’s best to use a Node version manager like nvm.

2. Set your OpenAI API key:

   For persistent access, you can add this line to your shell configuration file (e.g., ~/.bashrc or ~/.zshrc), or place it in a .env file in your project directory.

3. Start Codex:

However, you won’t need to do any of this for the course—everything is already set up for you in the course environment. The CLI is ready to use, so you can focus on learning and experimenting without worrying about installation or configuration steps.

When you launch Codex CLI by running codex in your terminal, you enter an interactive session. After you type your first instruction in plain English—such as “Write tests for date_utils.py”—Codex analyzes your request and responds with a proposed change. You’ll see a summary of what Codex plans to do, along with a preview of the suggested code or file modifications. The interface allows you to review the output before anything is applied to your files.

Codex uses the files in your current directory as context, so it’s important to be in the right folder when you start. This helps Codex understand your project and provide more accurate suggestions.

If Codex doesn’t understand your request, try rephrasing it or being more specific. For example:

• Vague:
  Fix my code
• Better:
  Fix the bug in the fetch_data function in api.py where it crashes on empty input

You can always ask Codex to explain its suggestions before you approve them. Codex will never change your files without showing you a preview first, so you remain in control of your codebase at all times.

While Codex is a powerful assistant, it’s not perfect. Sometimes its suggestions may need a bit of tweaking, and it’s always a good idea to review changes before applying them. If you ever get stuck or want to learn more, you can refer to the course Q&A or check the official documentation.

You’re now ready to start exploring Codex CLI in practice. The next steps in the course will guide you through real examples and hands-on exercises, so you can see how Codex fits into your own workflow.