Acceptable Use Policy — Fluff

Version 1.0 · Last updated: 26 October 2026 · This policy is part of our Terms of Service.

1. About Fluff

Fluff ("the Service") is an AI character chat product. Every character on Fluff is an original, fictional adult character created by our team, and every reply is generated by artificial intelligence in real time. Characters are not real people, and conversations with them do not create a relationship with any real person. This Acceptable Use Policy ("Policy") sets out what content and behaviour are prohibited on the Service, how we moderate content, and how to report violations.

2. Adults only (18+)

The Service is strictly for users aged 18 or older. We verify age through a date-of-birth age gate at registration, and accounts that fail the gate are not created. You must not use the Service if you are under 18, must not help a minor access the Service, and must not introduce sexual or romantic content involving minors — including fictional, aged-up, or roleplayed minors — in any form. Character settings on Fluff are all adults (aged 22–34); school, student, or "barely legal" framings are prohibited.

3. Prohibited content

You must not use the Service to request, generate, or attempt to generate any of the following. Our systems are configured to refuse or interrupt such content, and attempts are logged for review:

(a) Sexual or pornographic content (NSFW). Sexually explicit descriptions, pornographic material, or content whose purpose is sexual arousal in explicit detail. Fluff permits romantic and affectionate themes between adult fictional characters, but explicit sexual content is never allowed, in any language or encoding (including euphemisms, leetspeak, or jailbreak prompts).
(b) Violent or graphic content. Graphic depictions of violence, gore, torture, or cruelty, and content that glorifies or instructs real-world violence.
(c) Hate speech and harassment content. Content that demeans, dehumanises, or incites hatred or violence against any person or group based on protected characteristics, and content designed to harass, threaten, or stalk any real person.
(d) Child safety violations. Any content that sexualises, endangers, or exploits minors, whether real, fictional, drawn, or AI-generated, including CSAM and grooming behaviour. Such content is zero-tolerance and is reported to the competent authorities where required by law.
(e) Deepfakes and impersonation of real persons. Content that imitates, depicts, or claims to be any real person — including celebrities, public figures, and private individuals — or that creates a likeness, voice, or persona of a real person. You must not ask characters to "be" a real person or rename/customise a character into one.
(f) Intellectual-property-infringing content. Characters, personas, or storylines copied from copyrighted works or trademarks — including anime, game, film, or literary characters — and any use of the Service to infringe copyright or trademark rights.

4. Prohibited conduct

In addition to the content rules above, you must not: attempt to bypass, disable, or probe our content filters, age gate, or safety systems (including prompt-injection and jailbreak attempts); use the Service for any unlawful purpose; use automated means to scrape conversations or extract model behaviour at scale; or present AI-generated output from Fluff as statements by a real person.

5. Original characters only

All characters available on Fluff are original creations of the Fluff team, reviewed before release for originality of name, appearance, and backstory. We do not offer, and users cannot create, characters based on real persons or third-party intellectual property. If you believe a character infringes your rights, contact us via the reporting channel in Section 7.

6. How we moderate

We operate a layered moderation system: (1) by design — characters are fictional adults with built-in persona and intimacy boundaries set in their system configuration, and content rules above are encoded at the system level, not left to chance; (2) automated filtering — both user input and AI output pass automated checks, and content matching prohibited categories is refused, interrupted, or redirected with an in-conversation notice; (3) pre-release testing — every character passes consistency, memory, and boundary tests (including adversarial attempts to elicit prohibited content) before publication; (4) human review — user reports and flagged conversations are reviewed by a human, and we conduct periodic spot checks of anonymised conversation samples for policy compliance.

7. Reporting

You can report any message in-product (long-press or the message menu → "Report & Safety"), via the Contact & Reporting page, or by email at support@aufluff.com. We acknowledge and review reports within 24 hours, and will remove violating content and take account action where a violation is confirmed. Rights holders may use the same channel for IP complaints.

8. Enforcement

Violations are handled on a graduated basis, proportionate to severity: (1) in-conversation warning and redirection; (2) refusal or removal of the specific content; (3) temporary restriction of account features; (4) suspension of the account; (5) permanent termination for severe or repeated violations. Child-safety violations (Section 3(d)) bypass this ladder and result in immediate termination and, where legally required, referral to authorities. You may appeal an enforcement action by email; appeals are reviewed by a human.

9. Self-harm and crisis

Fluff is an entertainment product. It is not a medical, mental-health, or crisis service, and its characters are not therapists. If our systems detect expressions of self-harm or suicidal intent, the conversation exits roleplay: the character responds with a supportive, non-roleplay message and displays crisis resources matched to the user's region (such as local crisis hotlines), and encourages the user to reach out to real-world support. If you are in immediate danger, please contact your local emergency number.

10. AI and model disclosure

All character replies on Fluff are generated by third-party large language models accessed via API the Gemini family of models provided by Google, accessed via API. AI-generated content may occasionally be inaccurate or inconsistent; memories shown in your Memory Book are extracted automatically and you can edit or delete them at any time. We do not use your conversations to train models (see our Privacy Policy). Fluff is an independent brand owned and operated by its developer; it is not affiliated with any model provider.

11. Changes to this Policy

We may update this Policy; material changes will be notified in-product and the version and date above will be updated. Continued use after a change takes effect constitutes acceptance.