Developer Setup Guide

Your OS:

Claude Code

by Anthropic

An agentic coding assistant that runs in your terminal and VS Code. Reads your files, writes and edits code, runs commands, and explains its reasoning. Powered by Claude.

AccountClaude Pro / Max / Teams or API key

InstallNative installer (no Node.js needed) or npm

Docsdocs.anthropic.com

Codex CLI

by OpenAI

A terminal-based coding agent that works across your local files and codebase. Understands context, proposes edits, and can run code autonomously. Powered by GPT-5 Codex.

AccountOpenAI API key or ChatGPT subscription

InstallSingle install script (Mac/Win)

Docsgithub.com/openai/codex

1 · Install VS Code

Both tools + WSL2 step required

Download and install VS Code

Go to code.visualstudio.com and download the version for your operating system. Run the installer and accept all defaults.

Mac tip: After installing, open the Command Palette (Cmd+Shift+P) and run Shell Command: Install 'code' command in PATH so you can open projects from the terminal with code .

Windows: During installation, check "Add to PATH" — this lets you open VS Code from the terminal by typing code .. Also check "Register Code as an editor for supported file types."

Install the Python extension

Open VS Code. Press Cmd+Shift+XCtrl+Shift+X to open the Extensions panel. Search for Python (by Microsoft) and install it. This gives you syntax highlighting, linting, and notebook support.

Open the integrated terminal

Press Ctrl+` (backtick)Ctrl+` (backtick) to open a terminal panel inside VS Code. This is where you'll run all the commands in this guide. You can also open it via Terminal → New Terminal in the menu.

Everything you install in the terminal this session carries over. You don't need to re-install things each time you open VS Code.

Windows only: install WSL2

Both Claude Code and Codex CLI work best on Windows through WSL2 (Windows Subsystem for Linux). Open PowerShell as Administrator and run:

PowerShell (Administrator)

wsl --install

Restart your computer when prompted. After restarting, open Ubuntu from the Start menu and set a username and password when asked. Then in VS Code, install the WSL extension and use Remote → Connect to WSL — your terminal will now behave like a Linux system.

2 · Terminal basics

Both tools

Before installing anything else, get comfortable with a few essential commands. These work in the VS Code terminal. Windows users: run these inside your Ubuntu (WSL2) terminal, not PowerShell.

Command	What it does	Example
pwd	Print working directory — shows where you are	`/Users/you/projects`
ls	List files and folders in the current directory	`ls` or `ls -la` for details
cd	Change directory	`cd Desktop` · `cd ..` to go up
mkdir	Make a new folder	`mkdir my-project`
python3	Run a Python file or open the REPL	`python3 script.py`
pip install	Install a Python package	`pip install pandas scikit-learn`
git init	Start tracking a folder with Git	Run inside your project folder
git status	See which files have changed	Run this before every commit

Tip: Press the ↑ arrow key in the terminal to cycle through your previous commands. Press Tab to auto-complete file and folder names.

3 · Install Claude Code

Claude Code

Run the installer

The native installer is the easiest method — no Node.js required. Paste this into your terminal:

macOS terminal

curl -fsSL https://claude.ai/install.sh | sh

Windows PowerShell

irm https://claude.ai/install.ps1 | iex

WSL users: run the macOS/Linux command above inside your Ubuntu terminal instead.

The installer downloads the Claude Code binary, sets it up in your PATH, and enables automatic updates.

Verify the install

terminal

claude --version

You should see a version number. If you get "command not found", close and reopen the terminal and try again.

Authenticate

Run claude in your terminal. It will open a browser window to log in with your Anthropic account (Claude Pro, Max, or Teams). Complete the OAuth flow — you won't need to do this again.

Using an API key instead? Set it as an environment variable: export ANTHROPIC_API_KEY=your_key_here

Install the VS Code extension (optional but recommended)

In VS Code, open Extensions (Cmd+Shift+XCtrl+Shift+X), search for Claude Code, and install the official Anthropic extension. This adds a sidebar panel and lets you manage diffs visually rather than just in the terminal.

Your first session

Navigate to a project folder and run:

terminal

cd ~/Desktop/my-project
claude

Try asking: "What Python files are in this directory and what do they do?" Claude Code reads your actual files and responds in context.

4 · Install Codex CLI

Codex CLI

Run the installer

macOS terminal

curl -fsSL https://chatgpt.com/codex/install.sh | sh

Windows PowerShell

powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"

WSL users: use the macOS/Linux command inside Ubuntu instead.

Verify the install

terminal

codex --version

Set your OpenAI API key

You'll need an OpenAI account with API access, or a ChatGPT subscription. Set your key in the terminal:

macOS — add to your shell profile so it persists

echo 'export OPENAI_API_KEY="your_key_here"' >> ~/.zshrc
source ~/.zshrc

Windows (PowerShell) — set for current session

$env:OPENAI_API_KEY = "your_key_here"

To make this permanent on Windows, search "Environment Variables" in the Start menu and add OPENAI_API_KEY as a User variable.

Replace your_key_here with your actual key from platform.openai.com/api-keys.

Your first session

terminal

codex

Try: "Write a Python function that loads a CSV with pandas and prints summary statistics." Codex will propose code and ask for your confirmation before writing to disk.

5 · Using both tools: a quick comparison

Both tools

Both tools do similar things — the difference is mostly in how they authenticate and who makes them. Using both helps you develop your own sense of their strengths.

Claude Code — start a session

terminal

cd my-project
claude
> Analyze my data.py file and suggest improvements

Codex CLI — start a session

terminal

cd my-project
codex
> Analyze my data.py file and suggest improvements

Feature	Claude Code	Codex CLI
Model	Claude (Anthropic)	GPT-5 Codex (OpenAI)
Auth	OAuth via claude.ai or API key	OpenAI API key
VS Code integration	Official extension available	Via IDE integrations
Reads your files	Yes — full project context	Yes — full project context
Runs code	Yes, with your confirmation	Yes, with your confirmation
Proposes diffs	Yes	Yes

6 · Set up Git for your projects

Both tools

Both Claude Code and Codex work much better inside Git repositories. Git also gives you a safety net — if an AI makes a bad edit, you can roll back.

Configure Git with your name and email

terminal — run once

git config --global user.name "Your Name"
git config --global user.email "you@example.com"

Initialize a repo for each project

terminal

mkdir my-project
cd my-project
git init

Now run claude or codex inside this folder. Both tools will detect the repo automatically.

Commit before letting AI make changes

Get in the habit of committing your work before asking an AI to edit files. That way you can always compare or revert.

terminal

git add .
git commit -m "baseline before AI edit"

7 · Common issues

Both tools

"command not found: claude" or "command not found: codex"

Close your terminal, open a new one, and try again. The installer adds the command to your PATH, but the current session may not have reloaded it. If it still fails, check that the install script completed without errors.

Windows: "execution policy" error in PowerShell

Open PowerShell as Administrator and run Set-ExecutionPolicy RemoteSigned, then try the install command again.

API key not recognized

Make sure you've sourced your shell profile after adding the key. On Mac run source ~/.zshrc. On Windows, restart the terminal. You can also verify it's set with echo $OPENAI_API_KEY (Mac) or echo $env:OPENAI_API_KEY (Windows).

Still stuck?

Reach out via the CAID contact page with your OS, the exact command you ran, and the full error message.