NewsletterBlogLearnCompareTopicsGlossary

How Do You Download and Install the Codex CLI?

Download the Codex CLI with npm install -g @openai/codex. Requires Node.js 22+ and an OpenAI API key. Works on macOS, Linux, and Windows via WSL.

tools
ShareXLinkedIn

How Do You Download and Install the Codex CLI?

The Codex CLI is installed via npm with a single command: npm install -g @openai/codex. It requires Node.js 22 or higher and an OpenAI API key set as the OPENAI_API_KEY environment variable. The source code is open source and available at github.com/openai/codex.

Context

This question comes up because the Codex CLI is not a traditional installer download — there's no .exe, .dmg, or .pkg to click through. It's a Node.js package distributed through npm, which means the download and installation happen in one terminal command. That surprises developers who expect a binary release on a GitHub releases page.

The CLI works natively on macOS and Linux. Windows users need to install via WSL (Windows Subsystem for Linux) — a native Windows binary is not currently available. For a deeper look at what the Codex CLI does once installed, see OpenAI Codex: The Complete Guide to Cloud-Based AI Coding Agents.

Before you install, confirm your Node.js version is 22 or higher with node --version. Earlier versions will fail silently or throw compatibility errors. If you're evaluating Codex CLI against other terminal-based agents, the Codex CLI vs Claude Code comparison covers the key differences in approach and workflow.

If you work with VS Code, the Codex CLI integrates with the editor — see OpenAI Codex VS Code Extension for setup details. Open source maintainers also have access to a dedicated free tier covered in OpenAI Launches Codex for Open Source.

Practical Steps

macOS / Linux:

  1. Confirm Node.js 22+: node --version
  2. Install the CLI: npm install -g @openai/codex
  3. Set your API key: export OPENAI_API_KEY=sk-... (add to ~/.zshrc or ~/.bashrc to persist)
  4. Verify: codex --version

Windows:

  1. Install WSL 2 (search "Turn Windows features on or off" → Windows Subsystem for Linux)
  2. Open a WSL terminal and follow the macOS/Linux steps above
  3. Native Windows support is not available at time of writing

Once installed, run codex in any project directory to start an interactive session.

Want more AI insights? Subscribe to LoreAI for daily briefings.