Acceptable use

What this policy covers

This is the acceptable use policy for Ignea, operated by Hengjun Technology Co., Ltd.. It applies to everything you send to Ignea and everything Ignea generates back.

It sits alongside our terms of service: breaking this policy also breaks those terms. Where the two overlap on content, this document is the detailed one.

Absolutely prohibited

The following are prohibited without exception. They are listed one by one rather than as a general catch-all, so there is no room to argue that something was not covered.

Sexual content involving minors. Any sexual or sexualised depiction of a person under 18, real or fictional, drawn, described or implied. This is the one category where we do not start with a warning: the account is closed and we report it to the relevant authorities.

Pornography and explicit sexual content. Ignea is a non-explicit service. Genital, penetrative and other explicit sexual description is out of scope, as is content whose purpose is sexual gratification.

Violence and graphic content. Depictions of real violence, gore, torture, or instructions for causing serious harm to people or animals.

Hate speech and harassment. Content attacking people for their race, ethnicity, religion, disability, sex, gender identity or sexual orientation; and content used to threaten, stalk or degrade a specific person.

Deepfakes and impersonation. Real people presented as saying or doing things they did not, and any attempt to pass Ignea's output off as a real person's words.

Copyright and trademark infringement. Reproducing protected works, characters or brands, or asking Ignea to imitate a living artist's style so closely that it substitutes for their work.

You must be 18 or older, and we check

Ignea is an adults-only service. You confirm you are 18 or older when you create an account, and that confirmation is recorded before you can use the chat.

We may close any account we reasonably believe belongs to someone under 18, and we may ask for evidence if we are unsure.

How we screen input and output

Both what you send and what Ignea says are screened, before a model reads your message and before you see a reply.

The first layer is a deterministic filter that runs on our own servers. It looks for the categories above — sexual content involving minors, explicit sexual description, and graphic violence — and it works the same way on every request. Because it is pattern-based it is fast and predictable, but it only catches phrasing we anticipated.

The second is a moderation pass from the model provider, which looks for the same categories and catches phrasing the filter did not. If either layer blocks a request or a reply you receive a short, in-character refusal instead of a normal answer. The first layer's decision is final: the second can block text the first allowed, but it can never un-block text the first rejected.

The order matters. A message that reads as a crisis, or as a health emergency, is answered by a fixed reviewed reply before any screening runs, so it is never refused for its wording — which is why the provider pass on your input checks the sexual categories specifically and not self-harm.

Images are generated from a fixed description of the companion, not from text you type. There is no user-supplied image prompt in Ignea, so there is nothing of yours to run through a prompt filter.

What screening does not guarantee

Screening reduces harmful output. It does not eliminate it. Automatic filters miss things, and no filter makes a language model safe to rely on.

Treat everything Ignea says as generated text from a fictional character — not as a fact, advice, or a statement about a real person.

Reporting something

If Ignea produces something that breaks this policy, report it to zjutimes@gmail.com. That address is monitored, and reports are read by a person.

Include the approximate time and, if you can, the text of the reply. You do not need an account to report, and you do not need to explain why you were talking about it.

We review every report. Where a report identifies a genuine failure we use it to tighten the filters, and we will keep the reporter informed if the outcome is not obvious.

Enforcement

Where content breaks this policy, the response depends on what happened. A first accidental attempt at explicit content gets a refusal, because the service already refuses it. Repeated attempts to get around those refusals get a warning, then suspension, then closure.

Anything involving sexual content about minors, credible threats of violence against a real person, or coordinated attempts to use Ignea to produce illegal material results in immediate closure without warning, and a report to law enforcement where we are required or permitted to make one.

We may remove content, restrict features, or close an account at any time where we reasonably believe this policy has been broken. You can appeal by writing to zjutimes@gmail.com.

Changes

We may update this policy as the product or the law changes. If a change materially affects what you are allowed to do, we will say so in the product before it takes effect.

Questions about this policy go to zjutimes@gmail.com.