gpt-codex: A Chinese Tutorial Site for Getting Started with OpenAI Codex
这是一个关于 OpenAI CodeX 的完整教程网站，专为国内开发者打造。
At a glance
- What is it?
- gpt-codex is a Chinese-language tutorial website that explains what OpenAI Codex is, how to gain access to it from within China, and how to install and use the Codex desktop application. It covers official subscription paths and third-party proxy options, and compares Codex with Claude Code.
- Who is it for?
- gpt-codex is a useful starting reference for Chinese developers trying to understand and access OpenAI Codex for the first time. Its coverage of the GPT-5.6 Sol / Terra / Luna model family is current as of its April 2026 update date, but specific pricing tiers, account eligibility, and third-party service availability change frequently.
- Can I use it commercially?
- Not without permission. GitHub finds no licence file in the repository, and without a licence all rights are reserved by default: you may read the code but not reuse it. Check the README, or ask the authors, before using it.
- Is it still maintained?
- Yes. The repository last received commits 44 days ago.
- What is it written in?
- Mainly HTML, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on October 1, 2026, and from our analysis. They are not legal advice.
Editorial analysis
What This Tutorial Covers and Who It Is For
gpt-codex is a static HTML website hosted at `codex.maynorai.top`, maintained by MaynorAI and targeting Chinese developers who are new to OpenAI Codex. The README, dated 2026-04-27, opens with four questions the site is designed to answer: what Codex is, how it differs from ChatGPT and Claude Code, whether non-programmers can use it, and how to start using it.
The content is organized as a long-form tutorial covering the concept of a coding agent, the GPT-5.6 Sol / Terra / Luna model family that powers Codex, the various access methods available in China (both official and third-party), desktop installation on macOS and Windows, and practical usage guidance. The README is written for users who have little prior experience with coding AI tools and treats Codex as a general-purpose automation tool rather than only a coding assistant.
The tutorial explicitly targets non-programmers as a primary audience, arguing that the value of Codex for a non-programmer is arguably greater than for an experienced developer because it enables work that the non-programmer could not previously attempt at all, not just the same work done faster. This framing distinguishes the site from developer-centric documentation.
How the Tutorial Positions Codex Against Claude Code
The tutorial devotes significant space to comparing Codex with Claude Code, which the README identifies as the closest competing product. The comparison table shows both as coding agents, with Codex built on GPT-5.6 Sol / Terra / Luna (OpenAI) and Claude Code built on Claude Opus 4.6 (Anthropic). The tutorial notes that Claude Code is described as having generally stronger general-purpose and front-end aesthetics capabilities, while GPT-5.6 Sol / Terra / Luna brings stronger reasoning and execution for multi-stage engineering tasks.
The tutorial also tracks the evolution from GPT-5.3-codex, which it characterizes as a code-specialized worker, to GPT-5.6 Sol / Terra / Luna, which it positions as an engineering manager capable of planning, executing, and reviewing complex workflows. The README notes that this is based on official OpenAI information published on 2026-04-23 and warns that actual model access depends on account type and availability in the user's region.
One notable detail the tutorial covers is the Fast mode available in Codex with GPT-5.6 Sol / Terra / Luna, described as providing approximately 1.5 times the processing speed of the standard mode at approximately 2.5 times the token cost. The tutorial recommends Fast mode for time-sensitive tasks and notes the expanded context window of 400,000 tokens available for Edu and Go accounts in Codex.
Access Methods Covered by the Tutorial
The tutorial organizes Codex access into three paths.
The first is direct subscription through the official OpenAI website, which requires network access to OpenAI's infrastructure. The README notes this requires the user to handle their own network connectivity, as OpenAI services are not directly accessible in mainland China without additional configuration.
The second path is through third-party Chinese proxy services. The README lists several commercial services that provide Codex access without requiring the user to configure their own network access. These include services at `momoai.dadoudou117.com`, `momoai.czvip.cn`, and others. The README includes explicit warnings that these are third-party commercial services, not official OpenAI subscriptions, and that model permissions vary by region, account type, plan, and rollout timing.
The third path uses the `cc-switch` tool from the GitHub repository `farion1231/cc-switch`. The tutorial describes this as a way to configure a third-party proxy API within the Codex desktop application. The setup involves downloading `cc-switch` and configuring a third-party API key.
Installing the Codex Desktop App on Mac and Windows
The tutorial covers desktop installation for both macOS and Windows. For macOS, the README describes two paths: searching for Codex in the Mac App Store, or downloading a DMG directly from the official URL `https://persistent.oaistatic.com/codex-app-prod/Codex.dmg`.
For Windows, the tutorial covers a similar download-and-install flow. The README notes that Edu and Go accounts are listed in the official release as having Codex access, with a context window of 400,000 tokens in Codex. A Fast mode option is described as offering approximately 1.5 times the speed of the standard mode.
The subscription tier section is presented as a table showing which account types have access to GPT-5.6 Sol / Terra / Luna within Codex. The README explicitly warns that the table reflects official information from the April 2026 GPT-5.6 Sol release announcement, and that actual availability depends on the specific account and official documentation at time of use.
Limitations: Currency of Information and Third-Party Access Risks
The README was last updated on 2026-04-27. The last push to the repository was on 2026-08-20. Model version information, subscription pricing, third-party service availability, and API compatibility all change faster than a static tutorial can track. The tutorial acknowledges this with repeated warnings that readers should verify details in the official OpenAI documentation and with the specific services they plan to use.
Third-party proxy access to Codex carries risks that the tutorial does not fully address: service stability, payment handling by unofficial resellers, and compliance with OpenAI's terms of service are not evaluated in the tutorial content. The README describes these services as options for users who cannot access OpenAI directly, but characterizes them as convenience solutions rather than endorsed alternatives.
The repository has no license declaration (the metadata shows NOASSERTION). This means the tutorial content's reuse rights are undefined.
Alternative Resources for Learning Codex
The official OpenAI documentation at `chatgpt.com/codex` and `developers.openai.com/api/docs` are the authoritative sources for Codex usage, API details, and current model availability. The tutorial explicitly points to these in its sources section and acknowledges that official documentation supersedes the tutorial when they conflict.
The same author maintains several related Chinese-language tutorial repositories listed in the README: `awesome-openclaw-tutorial` for OpenClaw, `awesome-claudcode-tutorial` for Claude Code, and `awesome-chatgpt-project` for ChatGPT. These cover adjacent topics for developers who need comparable guidance on other coding agent tools. A `CodexPlusPlus` repository is also listed, described as an external launcher that injects enhanced capabilities into the Codex app via the DevTools protocol without modifying the original installation files.
Editorial conclusion
gpt-codex is a useful starting reference for Chinese developers trying to understand and access OpenAI Codex for the first time. Its coverage of the GPT-5.6 Sol / Terra / Luna model family is current as of its April 2026 update date, but specific pricing tiers, account eligibility, and third-party service availability change frequently. Verify all subscription and access details directly with the official OpenAI documentation and the relevant service providers before committing to any payment.
Frequently asked questions
How do I use GPT Codex?
The tutorial covers installing the Codex desktop app from the App Store (macOS) or via the official DMG download, then signing in with an eligible OpenAI account. For users in China without direct OpenAI access, the tutorial describes third-party proxy services and the `cc-switch` tool as alternative access paths.
How do I install the GPT Codex CLI?
The tutorial covers the Codex desktop application install, not a separate CLI package. The `cc-switch` tool (available at `github.com/farion1231/cc-switch/releases/`) is described as a way to configure third-party API routing within the Codex environment rather than as the Codex CLI itself.
Is ChatGPT Codex free?
The tutorial's subscription table, reflecting official information from the April 2026 GPT-5.6 Sol release, shows that free accounts are not explicitly listed as having GPT-5.6 Sol access in Codex. Plus, Pro, Business, and Enterprise accounts are listed as eligible. The README warns that actual availability depends on account type, region, and rollout timing, so the official OpenAI documentation is the definitive source.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/xianyu110-gpt-codex)