English · 简体中文 · Read Online · Learning Path · Entry Map · Configuration · Practice · Recipes · References · Roadmap
From the first hands-on session to real workflow adoption, CodexGuide helps people use Codex for development, creation, research, automation, and team collaboration. If this project saves you exploration time, a Star helps more people discover it.
CodexGuide is available at codexguide.ai.
The GitHub README is a quick project overview. For actual learning, the website provides fuller navigation, search, sidebars, screenshots, setup maps, learning paths, and hands-on examples. Key pages are marked with review dates whenever possible, so you can decide when to verify details against official OpenAI sources.
If you are new to Codex, start with the Learning Path. If you already know whether you want CLI, Desktop App, Cloud, or IDE workflows, start from the Entry Map.
Codex is evolving from a coding assistant into an AI workflow system across CLI, Cloud/Web, IDE extension, desktop app, mobile collaboration, browser control, and automation.
CodexGuide is a practical knowledge base for real tasks. It focuses on three questions:
- How to start: which entry point, task, and setting should beginners choose first.
- How to deliver: how to describe work clearly so Codex can read projects, edit files, run commands, and return checkable results.
- How to retain learning: how to turn one successful run into reusable templates, rules, cases, and safety boundaries.
The guide is mainly written in Chinese today, but its audience is broader than Chinese-speaking developers. It also covers creators, researchers, product and operations teams, technical writers, team leads, and non-developer workflows that can benefit from Codex.
- First-time Codex users: follow the Desktop App, subscription, settings, mobile collaboration, and first-task path to complete one full loop.
- Developers bringing Codex into projects: learn CLI, IDE, Git, tests, CI, AGENTS.md, sandboxing, and approvals.
- Creators and knowledge workers: use Codex for writing, slides, research organization, knowledge bases, browsers, and workflow automation.
- Team leads and tool builders: build team rules, task templates, permission boundaries, retrospectives, and portable case libraries.
- People choosing an entry point: compare Desktop App, CLI, Cloud, IDE, ChatGPT mobile, and plugin ecosystems.
Want to appear here? Add me on WeChat and include your product name plus a short project sponsorship note in the friend request.
| Sponsor | Description |
|---|---|
 |
Thanks to Unity2.ai for sponsoring this project. Unity2.ai is a high-performance AI model API relay platform for individual developers, teams, and enterprises. It has long served leading Chinese enterprises, handles more than 30 billion token calls per day, and supports 5000 RPM-level high concurrency. It supports balance-based billing, first top-up bonuses, bundled subscriptions, enterprise invoicing, and dedicated support. Register through this link to receive a $2 balance, then join the official community to receive another $10 balance, for up to $12 in free credits. |
 |
PayForChat supports ChatGPT and Claude membership activation and top-ups: pay directly with WeChat, no overseas credit card required, top up your own account through official channels, receive a full refund if it fails, and get Chinese support plus receipts. Click here to register and try it. |
 |
GetGPT Pro provides AI subscription activation and top-up services for ChatGPT, Claude, and other tools, with self-service ordering, fast delivery, and after-sales support. |
 |
Project sponsor. PPToken provides API relay and key distribution for ChatGPT, Claude, Gemini and other mainstream AI models, with low latency, high availability, pay-as-you-go billing, and flexible subscription plans. |
 |
Project sponsor. Ciyuan API aims to become a one-stop AI interface platform for developers, providing stable, low-latency, and highly available large model API services to make AI application development simpler. |
| Content | Description |
|---|---|
| Learning path | Installation, login, subscription, settings, mobile collaboration, and the first low-risk task |
| Entry map | How to choose between Desktop App, CLI, Cloud, IDE, ChatGPT, and integrations |
| Configuration | CLI options, config.toml, MCP, Skills, Subagents, and safety approvals |
| Workflow practice | Task design, verification, non-development workflows, and team playbooks |
| Recipes | PPT, Draw.io, browser, Obsidian, Feishu, Figma, Notion, CI fixes, and more |
| Official references | OpenAI source material, GitHub repositories, and key fact links |
Read the Learning Path, then continue with Desktop App Download And Installation, Plus / Pro Subscription, Desktop App Overview, and First Task.
Start from CLI Installation And Login, then read First Local Code Task, AGENTS.md, and Sandbox And Approvals.
Start with the Team Playbook, then add Configuration And Extensions, Security Administration, Troubleshooting, and the Recipe Library.
| Module | Use It For |
|---|---|
| Learning Path | Reading order from beginner use to team adoption |
| Entry Map | Choosing between CLI, Desktop App, Cloud/Web, IDE, and ChatGPT |
| Desktop App Download And Installation | Downloading, installing, logging in, and preparing Codex Desktop App |
| Mobile Collaboration | Following desktop tasks from the Codex entry in the ChatGPT mobile app |
| First Codex Task | Picking a low-risk task and completing a verifiable change |
| CLI Installation And Login | Installing Codex CLI locally and signing in |
| First Local Code Task | Running Codex CLI in a real repository |
| Configuration And Extensions | CLI options, config.toml, MCP, Skills, Subagents, and safety approvals |
| Practice | Task design, non-development workflows, and team practice |
| AGENTS.md | Writing project-level rules and collaboration boundaries for Codex |
| Sandbox And Approvals | Safety boundaries for files, commands, network, credentials, and production resources |
| Cloud, IDE, And Desktop App | Matching Codex entry points to work scenarios |
| Recipes | Reusable task templates and retrospectives for real projects |
| Official References | OpenAI source material, GitHub repositories, and key links |
CodexGuide
├─ guide # Practical guides from beginner use to team adoption
├─ platform # CLI, App, Cloud, IDE, and ChatGPT entry map
├─ configuration # CLI options, config.toml, MCP, Skills, safety approvals
├─ practice # Task design, non-development workflows, team practice
├─ recipes # Reusable real-world examples
├─ reference # Official references and fact sources
└─ community # Roadmap and contribution direction
Requirements:
- Node.js 22.12+ and lower than 25
- pnpm 10.33.0
pnpm install
pnpm devBuild the static site:
pnpm buildThe development server starts the VuePress documentation site. The production site is published at codexguide.ai.
- Official sources first: features, pricing, availability, and safety policies should be checked against OpenAI sources.
- Beginner friendly: onboarding chapters explain why a step matters and when to avoid it.
- Real task oriented: fewer abstract concepts, more repeatable task flows, inputs, outputs, and verification methods.
- Clear safety boundaries: file writes, command execution, network access, credentials, browser control, and computer control all need explicit risk framing.
- Reusable learning: successful runs should become AGENTS.md rules, templates, cases, retrospectives, and team norms.
Search 苍何 on WeChat or scan the QR code below. Reply with codex交流群 to join the community group.
This repository prioritizes official sources and marks important pages with review dates where possible. The current structure references:
- OpenAI Codex product page
- Codex in ChatGPT Help Center
- OpenAI Codex CLI Getting Started
- Codex cloud docs
- openai/codex GitHub repository
Contributions are welcome:
- Beginner-friendly rewrites.
- Reproducible real-world cases.
- Common errors and fixes.
- Team practices, templates, and workflows.
- Official documentation change tracking.
Please read the Contribution Guide. If you are not sure where to begin, start from the Roadmap or a good first issue.
This project is open-sourced under the MIT License. You may use, modify, distribute, and build on it as long as you keep the license notice.
CodexGuide is a community-maintained Codex practice knowledge base and is not an official OpenAI project. For time-sensitive details such as features, plans, pricing, availability, and safety policies, please refer to official OpenAI sources.