The Codex app on Linux

Which route fits your machine

A third option, Maestro, is a desktop app built on the CLI. It's covered at the end.

Install the Codex app

Check your processor first: uname -m prints x86_64 for x64, and aarch64 or arm64 for ARM64. Then download the matching package from OpenAI's Linux page.

Ubuntu or Debian

cd ~/Downloads
sudo apt install ./chatgpt_amd64.deb     # ARM64: chatgpt_arm64.deb

Open ChatGPT from your applications menu, or run chatgpt in a terminal, and sign in with your ChatGPT account to work with projects, local files and Codex. The package adds OpenAI's signed repository, so updates come with your normal upgrades:

sudo apt update
sudo apt install --only-upgrade chatgpt

Fedora

cd ~/Downloads
sudo dnf install ./chatgpt.x86_64.rpm    # ARM64: chatgpt.aarch64.rpm
sudo dnf upgrade --refresh chatgpt       # later, to update

Arch Linux

curl --proto '=https' --tlsv1.2 -fL -o install-arch.sh \
  https://persistent.oaistatic.com/codex-app-prod/linux/install-arch.sh
sudo bash install-arch.sh

The script adds OpenAI's signed repository and runs a full system upgrade, so it asks before it goes ahead. Later updates arrive with sudo pacman -Syu.

What the Linux preview doesn't do yet

• Other distributions. OpenAI supports only the ones listed above and says others may work but aren't formally supported.
• Computer Use, where the app operates other apps on your screen, is on macOS and Windows but not yet in the Linux preview.
• Native Wayland is experimental. The app uses XWayland when it's available. To try native Wayland, quit the app and run chatgpt --ozone-platform=wayland; floating windows, focus and keyboard shortcuts may not fully work yet.
• No AppImage, Flatpak or Snap. OpenAI publishes .deb and .rpm packages and the Arch script only.

On other distributions, or a server: the Codex CLI

The CLI is the same Codex in a terminal. It doesn't need a desktop session, so it also works over SSH.

curl -fsSL https://chatgpt.com/codex/install.sh | sh   # re-run it to update
cd ~/code/myapp
codex                                                  # choose "Sign in with ChatGPT"

npm install -g @openai/codex and brew install --cask codex work too. On a machine without a browser, run codex login --device-auth (in beta; turn on device code login in your ChatGPT security settings first), or sign in on another machine and copy ~/.codex/auth.json across. To use an API key instead of your ChatGPT plan:

printenv OPENAI_API_KEY | codex login --with-api-key

API-key usage is billed at standard API rates, and features that rely on ChatGPT workspace access or cloud services are limited or unavailable (OpenAI's auth docs).

Running several Codex tasks at once

In the app, choose Worktree under the composer before you send the prompt. Codex creates a git worktree from the branch you pick (in a detached HEAD by default), and Handoff moves the chat back to your main checkout when you want to test or commit there.

With the CLI, give each task its own worktree and terminal:

git worktree add ../myapp-auth -b feature/auth
cd ../myapp-auth && codex

The full routine, including merging the work back, is in how to run Codex agents in parallel.

Maestro: Codex on more distros, next to your other agents

Maestro is a free desktop app that gives the Codex CLI a window. It runs the codex you already installed and signed in, and every workspace is its own git worktree and branch, with diff review, pull requests, and your phone through Maestro Web. It ships as a single x86_64 AppImage that works across distributions, not only the ones on OpenAI's preview list, and it runs Claude Code, Cursor, OpenCode, Kimi Code and Grok Build in the same window. The AppImage needs the FUSE 2 library; see Maestro for Linux. More in Maestro as a Codex GUI.