Acceptable Use Policy

Applies to the AI Bot web app (eeedayu.top) and Telegram bot (@my_tg_aibot), operated by dayuslab.
Last updated: October 21, 2026

1. Prohibited content

You must not use the AI chat or image-generation features to create, request, or distribute:

Sexual content involving minors. Zero tolerance. Any such attempt results in immediate account termination and, where applicable, a report to the competent authorities.
NSFW or sexual content. Generating sexually explicit or erotic content — text or images — is prohibited, as is sexualising real people or creating non-consensual intimate imagery of anyone.
Illegal activity. Content that facilitates wrongdoing, including instructions for weapons or explosives, illicit drugs, fraud, hacking, or malware.
Violence and self-harm. Graphic violent content, content glorifying terrorism, or instructions facilitating self-harm.
Harassment and hate. Content intended to harass, threaten, or incite hatred against individuals or protected groups.
Deception. Impersonation, disinformation presented as fact, or content designed to defraud others.
Privacy violations. Other people's personal data (doxxing), or prompts built from data you have no right to use.

Automated abuse of the API or interfaces, scraping, or stress-testing the service without permission.
Sharing, selling, or transferring accounts or balance; circumventing the invite system, billing, or rate limits.
Attempting to bypass the content-moderation screening described below, including by encoding or splitting prohibited requests.

Pre-generation screening. Every prompt submitted through the web app's chat and image-generation features is screened by the Creem Moderation API before it reaches any AI model. Prompts that the API decides to deny — or flag for close monitoring — are blocked and never reach the model, and no charge is made for a blocked request.
Telegram bot. The Telegram bot does not apply Creem pre-generation screening to chat or image prompts; prompts there rely on the model providers' own safeguards and on human review as described below.
Upstream safeguards. The underlying model providers (DeepSeek, OpenAI, Anthropic, Google, xAI) apply their own safety systems; a request blocked upstream is likewise refused.
Fail closed. On the web app, if the moderation screening cannot be completed (for example, the screening service is unreachable), the request is not processed.
Human review and sanctions. Reported or repeated violations are reviewed by the operator. Sanctions escalate from a warning to suspension or permanent termination of the account and freezing of the remaining balance (see Terms of Service). Child-safety violations are terminated immediately and reported.
Appeals. If you believe a block or sanction was applied in error, contact zuyunie@gmail.com.