Codex Guide

On This Page8 sections

Editorial snapshot from September 2026. Prices, models and interface labels may change; this is a practical guide, not a live product feed.

OpenAI’s coding agent, available through CLI, IDE and cloud workflows.

Coding agent · OpenAI · macOS · Windows · Linux · Web

Suitable workflows

Developers evaluating a coding agent without making it their editor.

Limitations and billing

ChatGPT access and API billing are separate options. Features and limits depend on sign-in method.

Capabilities we verified

Agent workflow · CLI surface · MCP · Skills · Cloud agents

This is an evidence-based shortlist, not an exhaustive feature inventory. Omitted capabilities have not been checked for this record.

Pricing & usage

Paid from $8/mo

ChatGPT plan or API usage. The cost of metered usage or an external agent can be separate from the subscription.

Models & access

Model access varies by ChatGPT plan or API credentials.

Verified families: GPT.

Install and start the CLI

npm for the command below; standalone installers are also available for each desktop platform.

npm install -g @openai/codex

Open your project folder in a terminal, then run:

codex

Account and authentication

Run codex in your repository and complete the ChatGPT or API-key sign-in flow.

Project configuration and tools

Use AGENTS.md for project instructions and codex mcp to manage external tools. Review permissions before automation.

First task to try

Locate the project’s test command, explain the main modules, then review one small diff.

Script usage

This example asks for a project summary. Tool permissions still apply; a prompt is not a sandbox.

codex exec "Summarize the project structure without modifying files."

All AI CLI tools →

Claude Code — Anthropic’s coding agent for terminal and editor workflows.

Grok Build — Grok’s coding agent with a terminal interface, script mode and ACP integration.

Project evaluation checklist

Open a small, familiar project and confirm that your essential language tooling works.
Give the agent one bounded change with an observable acceptance criterion.
Review the diff, run the project’s checks and compare the result with your current workflow.
Inspect usage after the task. Estimate a month of similar work, including any separate model or agent bill.

Share this article

X Facebook LinkedIn Reddit Hacker News

Continue Reading

More articles connected to the same themes, protocols, and tools.

View all posts

TraeIDE Guides

Claude Code vs Codex

Kiro vs Google Antigravity

Cursor vs Zed

Referenced Tools

Browse entries that are adjacent to the topics covered in this article.

Explore directory

Floot MCP

AI App Builders / Prompt-to-App Tools

Floot MCP turns Claude, ChatGPT, Codex, Cursor, and other MCP clients into full-stack app builders connected to Floot's managed runtime. The assistant supplies the model and reasoning while Floot handles projects, backend services, previews, and publishing.

CodexBar

Developer Workflow Tools

CodexBar is a free, open-source usage monitor for AI coding tools that puts quota windows, reset times, balances, spend, and provider status into a macOS menu bar, Linux desktop app, and CLI. It is designed for developers juggling multiple coding-agent subscriptions rather than for writing code itself.

bb

AI IDEs / AI Code Editors

bb is an open-source, local-first agentic IDE built around orchestrating coding agents such as Claude Code, Codex, Cursor, Pi, OpenCode, Grok Build, and Hermes Agent. Its main differentiator is that the workbench itself is programmable: agents can create plugins, commands, skills, panels, and automations for bb.

Oh My Pi

Terminal / CLI Coding Agents

Oh My Pi (omp) is an open-source, terminal-first coding agent built from Pi and expanded into a batteries-included engineering harness with LSP, debugging, subagents, browser automation, model routing, and deep extensibility.

Cue