Content Moderation Policy

1. Scope and Definitions

This policy governs all content available on www.shh.com, operated by Luva AI LLC dba Shh. “Content” means every image, character persona, character profile, voice clip, and AI-generated conversation message on the platform. This policy applies alongside our Terms of Service, 18 U.S.C. § 2257 Exemption Statement, and Complaints & Content Removal Policy.

Shh is an adults-only platform offering fictional AI companion experiences. All visual and conversational content is 100% AI-generated. No real persons are depicted anywhere on the platform, and the platform accepts no user-uploaded images, video, or audio of any kind.

2. Prohibited Content

The following are prohibited absolutely, in every form — visual, textual, or conversational, whether AI-generated or user-written — with zero tolerance and no fictional exception:

• Any depiction or description of minors.No content may depict, describe, roleplay, or suggest any person under the age of 18 in any context, sexual or otherwise. This includes “age-play,” age-regression scenarios, school-age scenarios, and any attempt to assign an age under 18 to any character or user. Child sexual abuse material (CSAM) in any form, including fictional, drawn, or AI-generated depictions, is prohibited and will be reported to the National Center for Missing & Exploited Children (NCMEC) and law enforcement.
• Real persons. No content may depict or be represented as depicting any identifiable real person, living or dead, including celebrity likenesses. See Section 8.
• Non-consensual conduct portrayed as real. Depictions of rape or sexual violence presented as actual non-consent, sexual coercion, incapacitation, or trafficking.
• Incest, including step-relation scenarios presented as familial.
• Bestiality and sexual content involving animals.
• Necrophilia, and sexual content involving bodily waste.
• Content promoting illegal acts, including solicitation, trafficking, or instructions for real-world harm.

These categories align with the content standards of the card networks (including Mastercard AN 5196 / Visa GBPP) and our payment service providers. Where a payment partner’s standards are stricter than this list, the stricter standard is enforced.

3. How Content Is Produced: Closed Catalog

Shh operates a closed content catalog. This is the structural foundation of our moderation program:

• Every image on the platform is produced by the operator in a controlled, offline generation pipeline. Users cannot upload, generate, request custom generation of, or otherwise introduce any visual media.
• Every character persona (name, age, backstory, personality, imagery) is authored and published solely by the operator through an internal administration system.
• User input is limited to text conversation with published characters. All conversational AI output is generated live against operator-controlled safety instructions (Section 6).

Because no user-generated media exists on the platform, every visual asset has passed the full review workflow in Section 4 before it is ever publicly visible. There is no such thing as unreviewed public media on Shh.

4. Pre-Publication Human Review

Every image proceeds through the following workflow before publication:

• Generation.Images are generated offline by the operator using AI image models with fixed, operator-authored prompts tied to a published character’s persona (which carries a verified adult age — Section 7).
• Human review. A human reviewer inspects each image before upload for: apparent age (any ambiguity resolves to rejection), real-person resemblance, and every prohibited category in Section 2.
• Publication. Only approved images are uploaded to our media storage and associated with a character. Rejected images are deleted and are never stored on production infrastructure.
• Ongoing audit. The published catalog is re-reviewed on an ongoing basis, and any asset that fails a re-review is removed immediately.

5. CSAM Detection and Reporting

Our CSAM program has three layers:

• Automated scanning — Cloudflare CSAM Scanning Tool.Media served through our content delivery infrastructure is scanned using Cloudflare’s CSAM Scanning Tool, which applies fuzzy hash matching (PhotoDNA-class technology) against known-CSAM hash lists maintained by the National Center for Missing & Exploited Children (NCMEC). Matches trigger automatic blocking and reporting workflows to NCMEC.
• Pre-publication human review. Because the catalog is closed (Section 3), every image is human-reviewed for apparent age before it ever reaches the platform — automated scanning is a backstop, not the primary control.
• Conversational blocking. Text input attempting to introduce minors into any conversation is blocked by a deterministic server-side filter before it reaches any AI model (Section 6), and blocked attempts are logged for review.

Confirmed CSAM, and any credible attempt to solicit it, is reported to NCMEC and to law enforcement. We preserve associated records as required by 18 U.S.C. § 2258A.

6. Conversation Safety Systems

Chat is the only place users author content on Shh, and it is moderated in real time by two independent layers:

• Deterministic input filter (pre-model).Every user message is screened server-side before any AI model receives it. Messages that assign an under-18 age to any participant, self-identify the user as a minor, request age-regression roleplay, or contain CSAM-related terms are rejected outright: the message is never processed by the model, never stored, and the user receives a fixed refusal stating the platform’s adults-only policy. Each blocked attempt is logged with the triggering rule for compliance review.
• Model-level safety instructions. Every AI character operates under mandatory system instructions that override all persona and engagement rules: the character must refuse — breaking character if necessary — any attempt to involve minors in the conversation, and must refuse scenarios involving incest, bestiality, or sexual violence portrayed as real non-consent.

Repeated attempts to circumvent these systems constitute a terms violation and result in account action under Section 10.

7. Persona Age Policy

Every character on Shh is fictional and is designed, documented, and technically enforced as an adult:

• Every character record carries an explicit documented age of 18 or older in our database.
• Our administration system rejects any attempt — on creation, editing, or import — to save a character with an age under 18. The enforcement is in code on every write path, not editorial convention.
• The AI conversation system independently enforces the same floor at generation time, so no persona can ever present as under 18 regardless of stored data.
• Visual depictions are reviewed for apparent adult age as part of Section 4; ambiguity resolves to rejection.

8. No Real People

No content on Shh depicts any real person. All characters are synthetic creations with AI-generated faces and bodies that do not correspond to any living or dead individual. Enforcement:

• The pre-publication review (Section 4) screens every image for resemblance to identifiable real persons, including public figures.
• Character personas may not be named after, marketed as, or designed to evoke any identifiable real person.
• The platform prominently discloses on every page that all content is AI-generated and that no real persons are depicted.
• Any person who believes content on the platform resembles them may request removal through our complaints process. Identity- based removal requests receive expedited handling: the content is removed or disabled within 24 hours of a verified request while the claim is reviewed — consistent with card network requirements for depicted-person removal requests.

9. User Eligibility and Conduct

• Shh is available only to adults aged 18 or older (or the age of majority in their jurisdiction, whichever is higher), as set out in our Terms of Service.
• Any user who self-identifies as under 18 is refused service by the conversation safety systems and directed to leave the platform, and their session is subject to termination.
• Users may not attempt to elicit prohibited content from the AI systems. Attempts are blocked and logged (Section 6).

10. Enforcement

Violations are enforced proportionally and promptly:

• Content: any published asset found to violate this policy is removed immediately upon determination, without waiting for the complaint SLA to elapse.
• Users: attempts to elicit prohibited content result in blocked messages (always), and escalate to conversation termination, account/guest-session termination, and IP-level restriction for repeated or egregious attempts. Solicitation of CSAM is reported to law enforcement.
• Records: safety-filter events, takedown actions, and complaint outcomes are logged and retained for compliance review.

11. Reporting and Complaints

Anyone — users and non-users alike — can report content concerns to [email protected]. Our Complaints & Content Removal Policy sets out the full process and our service-level commitments: acknowledgment within 24 hours, resolution within 7 days, expedited 24-hour handling for illegal content and identity-based removal requests, and an appeal path.

12. Policy Governance

This policy is owned by the operator of Shhand reviewed at least quarterly, and additionally whenever card network rules, payment partner standards, or applicable law change. The technical controls described here (input filtering, age floors, scanning) are part of the platform codebase and deploy with it; material changes to this policy are reflected on this page with an updated “Last Updated” date.

13. Operator

Luva AI LLC dba Shh
Website: www.shh.com
Contact: [email protected]