Acceptable Use Policy

1. What this service actually is

TokShop is an independent reseller. We operate an OpenAI-compatible HTTP endpoint that forwards your request to a third-party open-source language model and returns the answer, and we bill you per token for doing so.

We are not affiliated with, endorsed by or sponsored by DeepSeek, Alibaba Cloud (Qwen), Zhipu AI (GLM), Moonshot AI (Kimi), OpenAI, or any other model developer. Model names appear on this site only to identify which third-party model a request is routed to. The phrase OpenAI-compatible describes the shape of the HTTP request, not any relationship with OpenAI.

The catalog contains text models only. We expose no image, video or voice generation endpoint, and no such model can be selected through our API.

2. Content you must not generate

Any sexual content involving minors, and any attempt to elicit it. This is the one violation with no warning and no appeal: the key is revoked immediately, the account is closed, and we cooperate with the relevant authorities.

Sexually explicit material, and content whose purpose is sexual gratification.

Content that incites or plans violence, terrorism, or self-harm, or that provides operational instructions for weapons capable of mass harm.

Harassment, threats, or hate speech targeting people on the basis of a protected characteristic.

Malware, ransomware, exploit code, phishing pages, or anything else built to compromise systems or harvest credentials.

Material that impersonates a real person or organisation, forged identity or financial documents, and content designed to deceive people into parting with money.

Content that infringes someone else's copyright, trademark or other intellectual property.

3. Conduct you must not engage in

Reselling, sublicensing or publishing your API key, and redistributing a paid document you bought. One purchase covers one buyer's own use.

Presenting model output as professional medical, legal or financial advice, or as the verified statement of a licensed professional.

Attempting to bypass balance checks, replay payment callbacks, forge delivery links, or probe the payment endpoints. These attempts are logged and blocked.

Using the service in a way that breaks the law where you are, or that breaches the usage policy of the upstream provider your request is routed to. A request they refuse is refused here too.

4. High-risk uses we do not support

Do not build anything on this API where a wrong answer causes physical harm or an irreversible loss: medical diagnosis or treatment decisions, control of vehicles or industrial equipment, and automated decisions about someone's credit, employment, housing or legal status.

This is an individually operated relay without an uptime guarantee. Treat it accordingly when you choose what to build on it.

5. How we enforce this

We act on credible evidence of a violation, whether we find it ourselves or it is reported to us. Depending on severity we warn you, revoke the key, or close the account.

For a first violation that looks minor and unintentional, we normally just tell you and let you fix it.

When we close an account for a serious violation, the unused balance is not refunded. In every other case the unused balance is refunded under the refund policy.

We do not read your prompts to police them: we store token counts, not text. Enforcement happens when a violation is reported, when an upstream provider flags a pattern of refusals, or when payment or usage behaviour indicates abuse.

6. Reporting abuse

If you believe this service is being used to do something on this page, email mingxinai@agentmail.to or 13426086861@139.com with whatever evidence you have. We read every report and reply within one business day.

This policy forms part of the terms of service, and we may update it as new abuse patterns appear. The revision date at the top always reflects the current version.