1. Purpose
This Policy describes how Velvet Room proposes to receive, assess, investigate, enforce, document, and review safety and policy matters. It is designed to accompany the Terms of Service and Community Guidelines and to align enforcement with the Service's privacy, persona, access, location, and value-bearing architecture.
2. Governing Principles
- Safety and consent: protect users from credible harm, coercion, exploitation, stalking, fraud, and serious abuse.
- Proportionality: match enforcement to severity, context, intent, repetition, impact, and available evidence.
- Privacy by discipline: investigate only what is reasonably necessary and do not treat moderation access as unrestricted product access.
- Account integrity: persona separation protects contextual identity from public enumeration, but does not allow an account to evade enforcement by switching personas.
- Evidence over assumption: distinguish verified facts, credible indicators, context, and uncertainty.
- No wealth-status favoritism: payment history, gifts, token/value holdings, subscription status, or creator/business status do not buy immunity from safety rules.
- Auditability: significant enforcement decisions should be traceable to policy version, evidence, actor, time, and outcome.
3. Current Moderation Authority - Source-Backed
The Legal/Policy Authority Audit identified an authenticated admin moderation route protected by JWT authentication and an admin-role requirement. That route expressly allows admins to view gated media URLs for moderation and states that message body content is not logged by that admin moderation surface.
The audit also identified an account-level isBanned field used in profile availability checks, a Channel status field capable of representing banned/paused states, and event/audit hooks in the moderation route. These are current technical signals; they do not prove that every enforcement action described below is already implemented. The audit additionally identified a WhisperDiscipline model containing screenshotSuspensionCount and suspendedUntil fields, which supports an existing screenshot-discipline concept but does not prove platform-wide capture detection or a final escalation policy.
4. How a Safety Matter May Begin
A safety or policy matter may be initiated by a user report where a verified reporting tool exists, authorized moderator review, security/abuse signals, a valid legal notice, a payment/fraud dispute, or information that reasonably indicates a potential violation. This draft does not claim a universal report endpoint until the live core workflow is verified.
5. Triage Framework
5.1 Critical / Immediate Risk
Examples include credible imminent threats, child sexual exploitation, trafficking, severe stalking/location abuse, account compromise creating immediate danger, or other circumstances requiring urgent safety action. Velvet Room may take temporary protective action before a full review where delay would materially increase risk.
5.2 High-Risk Abuse
Examples include non-consensual intimate material, extortion, repeated stalking or harassment, serious impersonation/fraud, coordinated abuse, dangerous attempts to defeat private access, or repeated evasion after prior enforcement. Repeated or deliberate capture of private/gated sexual content, intimate communications, or other high-sensitivity material in violation of the no-screenshot rule may also be treated as high-risk abuse.
5.3 Standard Policy Matters
Examples include spam, lower-severity harassment, misleading promotion, first-time boundary violations, copyright/policy disputes, or other conduct that ordinarily allows time for evidence review and notice. A first or lower-risk screenshot/capture violation may enter this category where there is no evidence of exploitation, redistribution, stalking, extortion, or evasion.
These triage labels are policy design categories introduced by this draft; they are not represented as existing database states.
6. Investigation and Evidence
- Review only evidence reasonably relevant to the reported or detected issue.
- Preserve context needed to distinguish satire, criticism, documentation, consent, coercion, impersonation, fraud, and credible threats.
- Treat restricted/gated media as reviewable by authorized moderation personnel where current authority permits it; access should be role-controlled and auditable.
- Do not state or imply that moderators routinely read private message bodies where the current moderation implementation does not log those bodies.
- Preserve evidence where reasonably necessary for security, dispute handling, enforcement, legal obligations, or an appeal, subject to the final retention matrix.
- Where screenshot/capture-discipline signals are technically available, use them as one piece of evidence rather than as infallible proof; capture detection can be device- and platform-dependent.
- Consensual adult explicit content located in a properly authorized private/gated adult surface is not a violation merely because it is explicit. The same content appearing on a public/discovery surface may be removed or gated, while non-consensual, exploitative, underage, or unlawful content remains prohibited everywhere.
- Avoid exposing sibling personas or sensitive location information to other users merely because it is relevant internally to an investigation.
7. Available Enforcement Actions
Depending on severity, confidence, context, repetition, legal obligations, and technical capability, Velvet Room may use one or more of the following actions. Not every action below is yet proven to be implemented in production and publication requires an engineering enforcement matrix.
- No action / close matter where evidence does not establish a violation.
- Educational notice or warning.
- Content visibility restriction, gating, removal, or disablement where supported.
- Restriction of a feature, interaction type, invitation/access mechanism, Radar/location capability, session capability, or other surface where supported.
- Screenshot/capture discipline, including warning, temporary restriction, or suspension where supported and proportionate.
- Persona or Channel restriction where supported.
- Temporary account suspension.
- Account ban/termination for severe, repeated, evasive, fraudulent, exploitative, or otherwise serious violations.
- Preservation or disclosure required by valid legal process or a genuine emergency, subject to applicable law.
8. Account-Level and Persona-Level Enforcement
A violation may originate from a specific Social, Friends, or Work persona, but serious enforcement may apply to the root account because the personas are extensions of one account. Velvet Room should not reveal hidden sibling personas to the public as part of enforcement notices or user-facing actions unless disclosure is lawful and necessary.
Creating or using another persona, account, device, invite, or private link to evade a valid restriction is itself a policy violation and may increase the severity of enforcement.
9. 18+ Eligibility Enforcement
If Velvet Room reasonably determines that an account is controlled by or being used for a person under 18, it may restrict, suspend, or terminate access under the Terms and applicable law. The launch product must provide a real server-side age-eligibility path; a client-only statement is insufficient.
Any evidence involving possible sexual exploitation of a minor must be handled under the highest-priority safety path and applicable legal obligations. This policy does not publish operational details that could weaken protective procedures.
10. Radar and Location Protective Actions
Because location and proximity information can create physical safety risk, credible misuse may justify fast restriction of Radar/location capability, access grants, proximity visibility, or account access while a matter is reviewed. Location features are not emergency services, and Velvet Room does not promise continuous monitoring or real-time intervention.
11. Notice of Enforcement
Where legally permitted, technically feasible, and not likely to create additional safety risk, the affected account should receive notice that identifies the action taken, a plain-language reason category, the policy section implicated, whether the action is temporary or permanent, and whether an appeal is available.
Notice may be delayed, limited, or omitted where necessary to protect another person, preserve an investigation, comply with law, prevent evasion, protect system security, or respond to an emergency.
12. Appeals - PROPOSED FRAMEWORK / IMPLEMENTATION REQUIRED
The Legal/Policy Authority Audit did not establish a complete canonical Velvet Room appeals workflow. The following is the required launch design, not a claim that it already exists.
- Eligible enforcement notices should provide a verified appeal path and state the applicable appeal deadline.
- An appeal should allow the user to identify the decision, explain relevant context, and submit information reasonably necessary to challenge the action.
- Appeals should be reviewed by a person or process reasonably independent from the original decision where practical, especially for account termination or high-impact actions.
- The reviewer may uphold, modify, reverse, or remand the action for further investigation.
- Velvet Room may decline repetitive, abusive, fraudulent, or substantially duplicate appeals after a final decision.
- A final appeal deadline, review target, and retention period must be locked before publication; this draft does not invent those timelines.
13. Repeat Violations, Evasion, and Severity Escalation
Repeated violations, coordinated abuse, retaliation, evasion, use of alternate personas/accounts/devices to bypass restrictions, or conduct showing deliberate exploitation of safety/access controls may result in stronger action than an isolated first incident. A single severe incident may justify immediate suspension or termination without prior warnings.
14. Abuse of Reporting and Enforcement Systems
- False or retaliatory reporting may itself violate the Community Guidelines.
- Threatening to report someone in order to obtain money, access, content, sexual interaction, or another benefit is prohibited.
- Attempts to fabricate evidence, coordinate mass false reports, bribe or impersonate moderators, or interfere with investigations may result in enforcement.
15. Emergency, Law-Enforcement, and Preservation Requests
Velvet Room may preserve or disclose information when reasonably necessary to comply with valid legal process, applicable law, or a genuine emergency request, subject to legal and operational safeguards. Velvet Room is not an emergency response service and should not promise 24/7 human moderation unless that capability is actually staffed and verified.
16. Records, Auditability, and Retention
Significant moderation actions should record, where technically implemented, the relevant account/resource identifiers, action, reason category, policy version, actor or automated authority, timestamps, evidence references, and appeal outcome. The current admin moderation route includes an audit-safe event hook, supporting the principle of traceable enforcement.
Exact retention periods for moderation evidence, safety reports, notices, and appeals are blocked until the platform adopts the final data-retention matrix. Records should not be kept indefinitely merely because they relate to moderation.
17. Transparency and Privacy Limits
Velvet Room may provide high-level transparency about enforcement practices when accurate and useful, but should not publish security-sensitive methods, expose reporters, reveal hidden persona relationships, disclose private location, or provide details that materially enable evasion or retaliation.
18. Engineering Requirements Before Publication
- Verify and document the canonical user-report submission path for core Velvet Room.
- Implement and verify the authenticated appeals workflow before promising appeals.
- Create an enforcement-action matrix mapping each public policy remedy to a real backend authority.
- Define moderation reason codes and policy-version references.
- Create user-facing notice delivery and status tracking.
- Lock retention periods for reports, evidence, actions, notices, and appeals.
- Verify age-gate enforcement and the operational under-18 handling path.
- Verify how reported private messages, private Rooms, Radar/location evidence, and gated media can be accessed lawfully and minimally for review.
- Standardize the platform-wide no-screenshot rule, capture-attempt event semantics, user notice, evidence standard, escalation schedule, and appeal treatment. Do not claim universal screenshot blocking unless technically proven.
- Verify adult-content classification and gating across upload, preview, Feed, Profile Preview, Channel, Radar/discovery, search/recommendation, moderation, and access-control surfaces.
- Wire canonical /safety and/or /enforcement-and-appeals public routes and versioned legal publication.
19. Operator, Contact, and Publication Gate
Operator: GGIRL Technologies LLC, New Hampshire, United States.
Safety/report contact: [SAFETY CONTACT / SUPPORT PATH] | Legal contact: [LEGAL EMAIL] | Legal notice address: [LEGAL NOTICE ADDRESS]