Codex Remote

You don't have to be at your keyboard for your code to get written.

In 2026, OpenAI put Codex in your pocket: kick off a task from your Mac, walk to a coffee shop, and review the diff, approve the command, and ship - from your phone. Mobile is fast becoming table stakes for any serious coding agent. But no book has shown practicing engineers how to actually run real engineering work remotely, safely, and without blowing your token budget. This is that book.

Codex Remote is the definitive, hands-on guide to OpenAI's coding agent, built around one idea reshaping how senior engineers work: your phone is the control plane. You're not typing code on a tiny screen - you're supervising a fleet of agents from anywhere, and that demands a discipline this book teaches end to end.

What you'll learn:

• The full Codex surface - the CLI, IDE extension, Cloud, the ChatGPT mobile app, the SDK, and the GitHub Action - for intermediate-to-advanced engineers, current as of June 2026 and leading with the latest model.
• Safe autonomy you can trust - the sandbox, the approvals model, auto-review, and blast-radius control, grounded in real security stakes (prompt injection, OWASP's top LLM risk). This is the judgment layer an agent can't hand you.
• The async / remote operating model - queue vs. steer, plan vs. goal, approvals as cards on your phone, and how to start work you check in on later instead of sitting and watching.
• Configuration that moves the needle - AGENTS.md done right (and the common ways it backfires), MCP, skills, subagents, and parallel fan-out.
• An honest comparison - where Codex fits next to other agents, and how to choose by workflow, not hype.
• Running real work without going broke - practical token-budget discipline for the end of the AI-subsidy era.

This isn't a button-by-button recipe an agent could regurgitate. It's the operator's playbook for agentic engineering - the production discipline beyond "vibe coding," organized around supervising work you don't have to be at a keyboard for.

Who it's for: practicing software engineers, leads, and platform teams adopting AI coding agents who want depth, safety, and a remote-first workflow - not a beginner walkthrough.

Pinned to documentation as of June 2026. This is an independent, unofficial guide. OpenAI, Codex, and ChatGPT are trademarks of OpenAI; this book is not affiliated with, authorized, or endorsed by OpenAI.

Stop walking back to your desk every time the agent needs a decision. Start engineering from anywhere - safely.