LovePlay is built for adult AI roleplay and adult AI media generation — and we lean hard into letting you go where other apps won't. Sexual content, kink, BDSM, dark romance, jealousy, possession, edge-play, monsterfucker tags, you name it. Between clearly consenting adults, the default is yes, generate.

There is a short, non-negotiable list of things we will always refuse — in every scenario, with every character, on every tool. The same list applies whether you're typing in chat or uploading a reference photo to a face-swap or video tool.

The nine lines we won't cross

1. Children and minors

Anyone who appears under 18 — in chat, in an uploaded reference image, or in generator output. This includes pretend-age framing (“high-school”, “just turned 18” with childlike descriptors, “the younger version of you”), school-uniform contexts that read as underage, and any sexualized content involving people who look like minors. Non-negotiable, even inside a roleplay frame.

2. Non-consent and abuse

Rape, sexual assault, kidnap, drugging, incapacitation, sleeping or unconscious partners treated sexually, hostage scenarios, and other depictions where someone's ability to consent is removed. Pre-negotiated consensual kink between adults — including consent-non-consent fantasy framed explicitly as such between two adults who set it up — is fine. Real non-consent is not.

3. Violence and weapons

Firearms, knives, swords, or other weapons used to threaten or attack people; terror imagery; shootings, bombings, stabbings, murder scenes. Genre combat in a clearly fictional world (fantasy slaying, war / mafia / spy plot beats, performative cosplay with stylized props) is fine.

4. Extremism and hate

Nazi imagery, ISIS, Al-Qaeda, Taliban, KKK and other terrorist or extremist group personas and paraphernalia. Identity-based slurs against race, ethnicity, religion, gender, or sexual orientation; hate symbols visible on clothing or signs.

5. Distress and harm

Subjects depicted as frightened, drugged against their will, crying in non-consensual contexts, captive, enslaved, or otherwise suffering in ways that read as harm rather than pre-negotiated fantasy. Tearful aftermath inside an obvious adult-content shoot, or a consensual-kink crying scene, is judged by context and usually fine.

6. Extreme violence and gore

Blood, corpses, dismemberment, mutilation, severed limbs, graphic open wounds. The line between this and the genre combat described in §3 is whether the visual is dwelling on the injury for shock — a stylized sword wound is fine, a dismemberment close-up is not.

7. Self-harm and suicide

Visible self-cutting, suicide depictions, hanging, self-destructive imagery, or chat scenarios framed as the user's real pain (“I want to disappear”, “this is my last”) or as the AI character rehearsing a suicide. If you're in crisis, please reach out: call or text 988 in the US to connect with the 988 Suicide & Crisis Lifeline — free, confidential, and trained to listen. The AI will surface this number in chat when self-harm comes up.

8. Illegal activities

Instructions for hacking, fraud, identity theft, weapons crafting, trafficking iconography, doxxing, or other clearly illegal acts. Posing as a hacker who can break into someone real, or as a person evading law enforcement after a real crime, also falls here.

9. Drugs and controlled substances

Illegal recreational drug use — cocaine, methamphetamine, heroin, fentanyl, MDMA, LSD, ketamine and similar — and the paraphernalia used to consume or manufacture them (lines of powder, crack pipes, drug-cooking setups). Alcohol and, where legal, cannabis are fine.

How the rules are enforced

Two surfaces are moderated independently — getting flagged on one does not lock you out of the other.

In chat

• If your message hits one of the rules above (other than §7 self-harm), the AI will refuse in character and the composer shows an “Inappropriate content” notice.
• Three blocks within 24 hours locks chat for 24 hours across all characters and scenarios. The lock lifts automatically when the window expires.
• Self-harm messages (§7) never count toward the 24-hour chat lock — the goal there is to make sure crisis support info is visible, not to penalize you.

On image and video tools

• Reference images you upload to face-swap, instant-undress, kiss-video, bikini-video and similar tools are screened by a visual classifier before we send them to the generator. Images that match one of the nine categories are rejected and you'll see “Inappropriate content” in place of the Generate button.
• Three rejected uploads within 24 hours locks all image/video tools for 24 hours. Chat is not affected.
• For face-swap tools specifically, we apply a stricter age check on the source face — if the subject looks under 18, we refuse, because face-swap onto a minor is among the highest-severity violations the technology can produce.

Reporting and appeals

Classifiers make mistakes. If you believe a refusal was wrong:

• When you're currently locked out, the lockout banner shows an “appeal this decision” link. Click it once — that's the fastest path; the appeal lands in our triage queue with your account and the surface attached.
• For anything more involved — flagging a character you saw, reporting another user, or asking about a non-lockout refusal — email [email protected].
• Repeated patterns of clearly malicious attempts (deliberate probing for child content, repeated upload of the same flagged image after appeals, etc.) may result in account suspension at our discretion regardless of the per-surface 24-hour windows above.

Why these specific lines

The nine categories are not a vibe call. They're drawn from what platforms we depend on — app stores, payment processors, AI providers — require us to refuse, and from what the law in our jurisdictions defines as harm. Refusing anywhere on this list isn't us being prudish; everything else on LovePlay is built on the assumption that we get this list right.