Document risks and non-recommended use of multi-account ChatGPT/Codex pools #2154

Documentation problem type

Missing documentation

Documentation location

README.md; ChatGPT/Codex account-pool documentation; Codex Auth / account-management documentation

What is wrong or missing?

OpenCodex presents the ChatGPT / Codex account pool as a first-class feature, including multi-account management, quota-aware routing, thread affinity, automatic selection, and failover.

The project already includes general warnings that users should review provider Terms of Service and that provider-side enforcement is outside OpenCodex's control.

However, the documentation does not clearly call out the provider-policy risk specifically associated with multi-account ChatGPT / Codex pool usage.

In particular, users may not realize that using multiple accounts to bypass rate limits, usage limits, quotas, plan limits, or other provider restrictions can create Terms-of-Use or account-policy risk.

This matters because the account pool is presented as a normal supported feature, and the README currently describes it as "Pool ChatGPT accounts safely". Without a nearby clarification, users may reasonably interpret this as meaning that multi-account usage is generally endorsed or low-risk.

I am not claiming that simply having multiple accounts is itself prohibited. The concern is specifically about ways a multi-account pool can be used to circumvent provider limits or restrictions.

This is not a request to remove or restrict the account-pool feature.

What should the documentation explain instead?

The documentation should make the project's support and risk boundary explicit.

It should explain that:

• Multi-account ChatGPT / Codex pool usage may carry provider-policy risk, particularly when used to bypass rate limits, quotas, plan limits, or other restrictions.
• The pool is provided as a technical account-management, routing, and resilience feature, not as an endorsement of circumventing provider restrictions.
• Users are responsible for ensuring that each connected account and their use of the pool comply with the provider's current Terms of Use and account policies.
• OpenCodex does not encourage creating additional accounts for the purpose of bypassing provider limits.
• OpenCodex does not encourage sharing account credentials between people.
• Provider-side account restrictions, suspensions, bans, or other enforcement actions are outside OpenCodex's control and are the user's responsibility.
• Maintainers are not expected to provide support for enforcement actions caused by a user's account setup or usage pattern.

A short warning should be placed close to the README account-pool description and in the relevant account-pool / Codex Auth documentation.

Suggested wording or example

Multi-account ChatGPT / Codex pools may carry provider-policy risk.

OpenCodex provides account-pool functionality for account management, routing, and resilience. However, using multiple accounts to bypass rate limits, usage limits, quotas, plan limits, or other provider restrictions may violate the provider's Terms of Use or account policies.

Such use is not endorsed by the OpenCodex maintainers and is entirely at the user's own risk. OpenCodex does not encourage creating additional accounts or sharing credentials in order to circumvent provider limits.

Users are responsible for reviewing and complying with the provider's current terms and policies. Provider-side account restrictions, suspensions, bans, or other enforcement actions are outside OpenCodex's control.

Prefer a single account or official API access where practical.

Additional context or attachments

Note: This issue was originally drafted in Japanese. The English text includes translation and wording assistance from an LLM. The underlying concerns, intent, and proposed changes are my own.

OpenAI's current Terms of Use prohibit circumventing rate limits or restrictions. I am intentionally not claiming that multiple accounts by themselves are prohibited; the concern here is the use of an account pool in ways that circumvent upstream restrictions.

OpenAI Terms of Use:
https://openai.com/policies/terms-of-use/

This issue is intentionally limited to documentation and support-boundary clarification. Changes to account-pool routing or implementation are out of scope.

A separate feature/UX issue can be opened if maintainers want an in-dashboard warning when configuring multiple ChatGPT / Codex accounts.

Checks

• I searched existing documentation issues.
• No secrets or personal information are included.