# One vendor's funnel, with a disclosure in the second paragraph and a five-point checklist written against itself > A Chinese-language guide, published as a hand-built static site, explaining how to obtain a paid ChatGPT subscription tier in China through one named third-party vendor. Every purchase link carries referral tracking, prices are dated inside the document, and the safety checklist is the vendor's own. The underlying service is not affiliated with the vendor of the product being sold. **fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide** — ChatGPT Pro 国内怎么开通？2026年 Pro 5x（100 美元）/ 20x（200 美元）、支付宝、无海外卡、Codex 与订单查询指南 - Repository: https://github.com/fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide - Website: https://fangmumu111-bot.github.io/chatgpt-plus-pro-codex-cn-guide/ - Stars: 1,463 · Forks: 69 - Language: HTML - License: MIT - Published: 2026-09-14 · Updated: 2026-09-14 · Language: en - Canonical page: https://hysenlabs.com/projects/fangmumu111-bot-chatgpt-plus-pro-codex-cn-guide ## The disclosure is in the second paragraph and the entry points follow it The document is maintained by the operator of a commercial service, and it says so near the top rather than in a footer. The wording is specific: the operator provides an independent third-party service and has no affiliation, agency relationship, or authorization from the company whose product is being sold. That is a stronger and more specific claim than a generic disclaimer would be, and it appears before the purchase links rather than after them. After it come four entry points, one per subscription tier, plus a general page for people who have not decided. Each tier has its own product identifier, and each link carries tracking parameters: the source is named as this repository's host, the medium is referral, there is a campaign name that contains the repository's own name, and there is a content label naming which link it was. There is also a separate order-query page and a comparison page. I am not reproducing those addresses, and the reason is the same reason the document prints a disclosure: they are the commercial endpoint of this page. Everything a reader is sent to, they are sent to the same vendor, and the guide's structure makes that visible even before you notice the tracking. What the guide claims for itself is narrower than the marketing tone suggests. It says the service handles orders manually, that you can query the order after paying, and that once the order completes you verify the tier in your own account. It also says the vendor's own product pages are authoritative for whether an account qualifies, for the subscription period, and for refund terms. ## Three purchase routes, and the one with no upsell attached has the best advice The document lays out three ways to get a subscription, and the ordering is not accidental. Route one is the official web checkout: sign in to your own account, open the settings or plan page, check which tiers your account currently shows, pay with a method your own checkout supports, and confirm the plan name and billing state afterwards. The document says this route suits people who already have a stable overseas payment method and want to manage renewal and cancellation themselves. Route two is the app stores. If your account and your store region show a purchasable tier, you can buy in app. The advice attached to this route is the most operationally useful line in the whole document, and it is a warning rather than a pitch: subscriptions, renewals, cancellations and refunds for an in-app purchase are all handled by the store you originally bought from, so before changing purchase channels you should confirm the old subscription is not still renewing, or you will be charged twice. Route three is the vendor, and that is the one with the affiliate links. So the route with the most practical guidance in it is the one that earns the guide nothing. That is worth noting as a matter of reading strategy: the operational warnings are the parts written by someone who has watched the failure, and the pitch is the part written for conversion. The document also gives an order of operations for choosing: start by comparing against the cheaper existing tier, and if you hit limits repeatedly on real work, then compare the two higher tiers, and first rule out login problems, client problems, or a usage figure that is simply not updating. ## The five-point checklist is the most defensible thing in the repository Before paying, the document asks the reader to check five things, and they are aimed at the vendor's own service. The account being topped up must be your own account, and specifically not a shared account, a rented account, or a pre-made account. That distinction is the whole of the difference between buying a subscription for yourself and buying access to somebody else's login. The purpose of the account information being requested, and how it is submitted, has to be stated clearly, and the checklist names the three things that must never be asked for: a password, a verification code, and a recovery code. Those three are the ones that convert a top-up into an account handover. Before payment you should be able to see the tier, the amount, the stock, the delivery method, and the refund rules. After payment there should be an order you can query, and the document is pointed about this: not only a screenshot in a chat conversation. And after completion, the tier should be verifiable on the official plan page in your own account. Two supporting claims sit next to the checklist. An unsuccessful top-up is described as fully refunded after the order is verified, which makes the order record the thing that determines whether a refund happens. And the document says verification of the tier and its validity happens in the official settings, not in the vendor's interface, which means the vendor's claim is falsifiable by a page you control. A vendor writing a consumer checklist against its own checkout is not the norm. It does not remove the commercial incentive behind the page, but it does mean the failure modes were at least considered by someone. ## The tier names carry multipliers and the document tells you not to read them that way The three tiers are named with multipliers, and there is a comparison table whose columns are the three tiers and whose rows are the dimensions of the choice. The official monthly prices are stated plainly for all three: 100, 200 and 500 dollars a month. The vendor's own renminbi price is not stated for two of them; those rows point at per-tier product pages instead. Only the top tier carries a number inside the document, dated as priced on the third of October 2026. The row for usage is the interesting one, because it is a hedge. The multiplier is labelled as a reference drawn from earlier material, with a note that new orders should be verified against the current plan, and for the third tier there is no multiplier at all, just an instruction to check current usage rules. In other words, the names in the table are marketing that the table itself declines to rely on. Then the sentence that most of these comparisons lack: do not understand the twenty-times tier as a smarter model, and do not treat an individual subscription as a seat to be shared between people. Both errors are plausible for a reader who has only seen a tier name and a price. The row describing who each tier suits follows the same logic. The cheapest is for people who hit the ceiling on the lower tier, or who use the coding tool heavily, or who work with long documents. The middle is for one person working all day across several projects. The top is for people who need a specific model access path in the coding tool and the hosted work surface, which is the one row that names a capability rather than a volume. ## Two model families are rolling out to paying tiers and the guide says a tier is not the condition A second half of the document is about models rather than plans, and it is where the guide is most careful. Two model families are named, along with an existing top-tier one that is described as a model option inside the chat surface. The note for the two new ones is that they are being rolled out in stages to existing subscribers and higher tiers in the hosted work surface and the coding tool, that ordinary chat cannot select them yet, and that the programmatic interface uses separate model identifiers and separate billing. The instruction that follows is the important one: check the model selector in your own account, and do not treat a particular subscription tier as the exclusive purchase condition for those models. That is a direct contradiction of how the rest of the page is built, and it is correct. The page exists to sell tiers, and the honest position is that access to a model is controlled by the upstream provider's rollout, not by what you paid a reseller. If a vendor told you otherwise, that would be a reason to stop. There is also a technical note with a specific floor: the command line tool has a stated minimum version, and the guide covers a minimum request shape, how to read the errors, and a reusable code task description. That is the only part of the repository that would be useful to somebody who is not buying anything. ## Every claim carries its own date, except the purchase links The document does something most pages of this kind do not, which is print the date next to each kind of claim. There are three separate dates and they are not the same. The content date says the guide was updated on the third of October 2026, and lists what changed. The service date says the acceptance status was last verified on the thirtieth of September 2026, and attributes that check to the site owner by name. The price date is attached to the single renminbi figure that appears in the text. And a note says that pricing and the technical model material each keep their own verification dates, which is the correct treatment when a page mixes facts with different half-lives. The repository also has a changelog, although it has published no tagged releases at all, and the default branch was pushed two days before the content date the page claims. So the editorial content is dated honestly and per claim. The interesting part is what is not dated: the purchase links. Each one is a tracking link with a campaign name, and a campaign name has no version. Nothing in the document tells you when a given product identifier was created, when its price last changed, or whether the item it points at is the same item the comparison table describes. If the content date is the freshness signal a reader uses, it applies to the advice and not to the checkout. That is the one asymmetry in an otherwise careful page, and it is exactly the asymmetry that matters to someone about to pay. ## A hand-built static site, a disclosure file, and a stray file named like a hash The repository is a website rather than a program, and the file list describes how it is put together. There is a site configuration file, a layouts directory, a hand-written index page, a small script, two stylesheets, one of which is a set of design tokens, an assets directory, a favicon, a robots file, a sitemap, and a plain text file intended for language models to read. It is published to a pages address derived from the account name, and there is no static site generator dependency and no build manifest, so the site is assembled by hand and committed. Two documents at the root are doing credibility work. One is a disclosure document, and the page links to it specifically for the identity of the service provider and the standard of evidence behind its claims. The other is a sources document. There is also a changelog. Then there is a file at the root named with a thirty-two character hexadecimal string and a text extension. There is no plausible reading of that as something a person intended to create. It has the shape of a pasted identifier or a truncated output of a command, and it sits beside a licence file as though it were documentation. There is no test, no build configuration and no package manifest. For a page like this that is entirely reasonable. The stray file is the one entry that suggests something ran unattended and nobody looked at the result, which is a small thing in a repository whose whole subject is trusting a vendor with your account. ## Conclusion Read this as marketing with unusually good disclosure rather than as documentation, because that is what it is. Every purchase link routes to one reseller's checkout and carries tracking parameters identifying this repository as the source, so the guide's value to its author is measurable and the value to you is a comparison of three tiers and a checklist you could have written yourself. The two things worth taking from it are the dates, which are printed rather than implied, and the checklist, which asks whether the account is your own and whether a password is being requested, which are the questions that distinguish a subscription top-up from account resale. Before using any service like this, understand what you are buying. A reseller's ability to open a tier on your account is a commercial arrangement between you and them, and neither the guide nor the vendor can promise what the upstream provider does with the account afterwards. The guide's own advice is the right one to follow: check your current plan on the official page, compare the cheaper tier first, rule out usage-display and login problems before paying for headroom, and verify the result on the official page rather than on a screenshot in a chat window. ## FAQ ### What is chatgpt-plus-pro-codex-cn-guide? A Chinese-language guide, also published as a static site, about how to obtain a paid subscription tier through one named third-party vendor when you have no overseas payment card. It compares three monthly tiers, links to the vendor's per-tier purchase pages and an order-query page, and includes notes on model access and a technical guide for the command line tool. ### Does chatgpt-plus-pro-codex-cn-guide claim to be affiliated with OpenAI? No, and it says so explicitly. The document describes the vendor as an independent third-party service with no affiliation, agency relationship, or authorization from the company whose product is being sold. It also states that account eligibility, the subscription period, and refund terms are governed by the vendor's own product pages. ### What does the guide say about account safety? That the account being topped up should be your own, and not a shared, rented, or pre-made account. That the purpose and submission method of any account information should be stated clearly, and that a password, verification code, or recovery code should never be requested. After payment, the tier and its validity should be checked on the official plan page in your own account rather than on a screenshot in a chat conversation. ### Which subscription tiers does the guide compare? Three, at 100, 200 and 500 dollars a month by the official price. The vendor's own renminbi price is given on per-tier product pages for two of them, and the top tier's price is stated and dated inside the document. The guide explicitly says the multiplier in a tier name does not mean a smarter model, and that an individual subscription is not a seat to be shared. ### Where should a purchase be verified after payment? On the official plan page in your own account, once the order shows as complete. The document treats that as the acceptance step and says orders can be queried afterwards. It also warns that if a subscription was bought through an app store, the store handles renewal, cancellation, and refunds, so you should confirm the old subscription is not still renewing before changing channel. ## Sources - [fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide on GitHub](https://github.com/fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide) - [Issues](https://github.com/fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide/issues) - [License: MIT](https://github.com/fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide/blob/main/LICENSE) - [Project website](https://fangmumu111-bot.github.io/chatgpt-plus-pro-codex-cn-guide/) - [README](https://github.com/fangmumu111-bot/chatgpt-plus-pro-codex-cn-guide/blob/main/README.md) --- Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/fangmumu111-bot-chatgpt-plus-pro-codex-cn-guide