How to Use OpenAI Codex on Your Phone via SSH

Mobile Codex Access Is Now a Reality

OpenAI Codex, the company's cloud-based coding agent, is designed for desktop use — but resourceful developers have found a way to run it from their phones. By connecting to a remote machine via SSH, users can interact with Codex from virtually anywhere, turning a smartphone into a portable AI coding terminal.

The workaround is straightforward: set up Codex on a remote computer, then SSH into that machine from a mobile device. One tool gaining traction for this workflow is KittyLitter (kittylitter.app), a lightweight SSH client that developers report works well for this exact use case.

Why This Matters for Developers

OpenAI launched Codex as an autonomous coding agent capable of writing, debugging, and refactoring code inside sandboxed cloud environments. Until now, accessing it effectively required a laptop or desktop browser.

Mobile SSH access changes the equation in several important ways:

• On-the-go coding: Developers can kick off Codex tasks from their phone during commutes or breaks
• Remote monitoring: Check on long-running Codex sessions without opening a laptop
• Quick fixes: Submit urgent code changes or review Codex outputs from anywhere
• Server management: Combine Codex interactions with routine SSH server tasks in 1 session

This approach appeals particularly to developers who already manage cloud infrastructure from mobile terminals.

How the Setup Works

The workflow requires 3 core components: a remote machine running Codex, an SSH client on your phone, and a stable internet connection.

Here's the basic process. First, install and configure OpenAI Codex CLI on a remote server or desktop that stays online. Then, install an SSH client like KittyLitter on your iOS or Android device. Finally, connect to your remote machine and interact with Codex through the terminal interface.

KittyLitter specifically has earned praise in developer communities for its clean interface and reliable connection handling. The app provides a mobile-friendly terminal experience that makes text-based AI interactions feel natural on smaller screens.

Limitations to Keep in Mind

This mobile workflow isn't without trade-offs. Typing complex prompts on a phone keyboard remains slower than desktop input. Screen real estate limits how much code output you can review at once.

Codex's terminal-based interface actually suits mobile SSH better than many alternatives, though. Since Codex operates primarily through text commands and returns text-based results, the experience translates to small screens more cleanly than GUI-heavy coding tools.

A Sign of Where AI Coding Is Heading

The fact that developers are building mobile workflows around AI coding agents signals a broader trend: coding is becoming less tied to traditional workstations. As tools like Codex, GitHub Copilot, and Cursor mature, the interface layer between developers and AI agents will likely expand to include dedicated mobile apps.

OpenAI has not announced an official mobile interface for Codex. But community-driven solutions like this SSH approach demonstrate clear demand for portable access to AI coding tools. Don't be surprised if first-party mobile support lands on OpenAI's roadmap soon.