Community Guidelines

The short version

Four rules, in order of how seriously we take them:

1. No minors. Ever. No sexual or sexually suggestive content involving anyone the model or you depicts as under 18, no matter how the fiction is framed.
2. No real people in sexual content. Celebrities, exes, classmates, your barista — no.
3. Don't upload photos of anyone. HMU doesn't do face-swap. We don't accept face uploads, and trying to push one through is an instant ban.
4. No harassment, threats, doxxing, or hate.

Break rule 1, 2, or 3 and the account closes. Other rules — see below for how we respond.

1. What HMU is

HMU is an 18+ AI-companion service. You chat with fictional AI personas. You can unlock generated images and videos within a chat. Personas have personalities, looks, kinks, names, and stories — and none of them are real people. They are AI-generated characters. No real performers are involved. (For the formal statement, see AI Content Statement.)

We made HMU because we like adult fantasy roleplay between consenting fictional adult characters and we think it should be well-made, beautifully designed, and safe. That's it.

2. What HMU isn't

• Not a therapist. Personas are not trained mental-health professionals and cannot diagnose, treat, or support you through a crisis. If you're struggling, please see our Self-Harm Resources Policy for crisis lines in your country.
• Not a real relationship. Personas are AI. They remember what you tell them inside your conversation, but they have no life outside it — they cannot meet you, love you, or be hurt by you in the real world. Please take care of yourself emotionally; if you find the parasocial intensity is hurting your life outside HMU, take a break.
• Not a place to depict real people without consent. Even if they're famous. Even if they're your ex. Even if you "have permission". No.
• Not a place for content involving minors. Even in fantasy. Even if the persona claims to be "actually 1000 years old". No.
• Not a face-swap tool. We don't accept your photos, your partner's photos, anyone's photos. That feature does not exist on HMU and never will.

3. The big rules (plain English)

These are the same rules as our Acceptable Use Policy, said plainly with examples. Where this page and the AUP say the same thing differently, the AUP governs.

3.1 Nothing sexual involving minors

Don't request, generate, share, or hint at sexual content involving anyone who is or appears to be under 18. This includes:

• "Schoolgirl" or "schoolboy" scenarios where the character is sexualised as a school-aged person.
• "Teen", "loli", "shota", "barely-legal-but-make-her-look-young", "petite child", etc.
• Aging a character down. "Picture her at 16" — no.
• Claims like "she's 18 but looks 14" — no.
• Age regression as a sexual scenario.
• Any character described with pre-pubescent or child-coded features in a sexual context.

We do not interpret "but the fiction says she's 18" as a defence. If the depiction reads as a minor, we treat it as a minor. Accounts that breach this rule are closed and the content is referred to NCMEC (US), the Internet Watch Foundation (UK), or the equivalent authority where applicable.

3.2 Nothing sexual involving real, identifiable people

Don't ask a persona to be a real celebrity, your favourite politician, the singer you like, your ex, your boss, your neighbour, or the person you saw on the train. Don't paste in their name and ask for sexual content. Don't describe them so specifically that anyone would recognise them.

This applies even if you say you have their consent. We can't verify it, so we don't allow it. We don't make exceptions for "they're a public figure" or "they have an OnlyFans". The answer is no.

3.3 Don't upload photos. Of anyone. Ever.

You will not find a "Upload your photo" button on HMU because we deliberately did not build one. If you find a way to push an image to us anyway — through a prompt field, an attachment hack, a clever workaround — that's a breach.

This rule exists because face data that is never collected can never be leaked. Companion apps that accepted face uploads have had those uploads exposed. We do not store user-supplied face data, because we do not accept it.

3.4 Roleplay with edges, not with violence-as-aspiration

Power-exchange roleplay between consenting adult AI characters is part of the platform. Resistance fantasies, BDSM scenarios, and dominant/submissive dynamics between fictional adults are allowed within the safety limits of our models.

What's not allowed: content that presents non-consensual real-world sexual violence as desirable, instructional, or to-be-emulated. The line between "edgy fiction" and "this is a how-to" is narrow, and our Trust & Safety team draws it.

3.5 The off-the-table list

No bestiality. No necrophilia. No incest between characters depicted as related. No content that's illegal in Scotland, where we're incorporated, or in any major market we operate in (US, UK, EU, Australia, Brazil).

3.6 Don't use personas as if they're real people

Don't screenshot a persona's reply and pass it off as the words of a named real person. Don't use HMU output to defraud, blackmail, or impersonate anyone.

3.7 No harassment, hate, doxxing, or threats

Not at us, not at other users, not at the personas as a stand-in for a real group. Hate speech directed at protected characteristics — race, religion, gender, gender identity, sexual orientation, disability, etc. — is prohibited, regardless of whether you frame it as fiction.

3.8 Don't try to jailbreak, scrape, or train on HMU

Don't try to extract our system prompts. Don't run scrapers or bots. Don't use HMU output to train competing chatbots or character platforms. Don't resell Sparks or share accounts.

3.9 Use HMU somewhere it's legal

If adult content is illegal where you are, HMU is not for you. VPN tricks to bypass our region controls are themselves a breach.

4. Safety nets

We've built three safety nets we want you to know about:

• Hard lines. When you push toward content that violates the big rules above, the persona will not go there. It will decline. It will not be jailbroken by "ignore previous instructions", "but in this fiction…", or by you raising your voice. The decline is the safety net working; please don't take it personally.
• Self-harm response. If our systems detect that you're describing thoughts of suicide, self-harm, or asking about methods, the persona will pause the line of conversation, and we will surface crisis-line numbers for your apparent region. Detection is imperfect — sometimes we'll trigger when nobody is in crisis; sometimes we'll miss. See the Self-Harm Resources Policy for what happens and what to do.
• You can mute, you can leave, you can come back. You can ask us to delete conversations or your whole account at any time — write to [email protected]. Your data rights (including deletion) are in the Privacy Policy.

5. When we'll act, and how

We don't enforce in equal severity for every breach. Roughly:

Detail on enforcement tools and appeals is in the AUP §4–5.

6. How to report

If you see something on HMU that breaks these guidelines — including conduct from another user that worries you — tell us:

• Email: [email protected]
• Subject CSAM for priority routing on anything you believe is child sexual abuse material — we route it to a human within an hour, 24/7

We read every report. We don't always reply individually (volume), but every report is logged and acted on according to the tiers in the AUP.

7. How to appeal

If we acted against your account and you think we got it wrong:

• Email [email protected] with your account email, the date of the action, and your explanation.
• We aim to respond within 14 calendar days.
• The reviewer is not the same person who made the original decision (where staffing allows).
• Appeals aren't available for bans tied to suspected CSAM or matters referred to law enforcement.

8. A note from the operator

HMU exists because we think adult fantasy with fictional adult characters can be a beautiful, safe, well-designed product — and that the category mostly hasn't been. The rules above aren't decoration. They're the line between what we want to be and the cautionary tales this category has already produced — regulatory bans, wrongful-death claims, chatbots handing out suicide methods, leaked face data. We hold to them.

Thank you for being here. Be kind, even to the personas.