Content & Safety Policy

Effective Date: March 4, 2026
Last Updated: July 13, 2026

1. Our Commitment to Safety

Safety is foundational to Petra's Villa — not an afterthought. We are committed to providing a creative, immersive, and secure environment for adult players to explore AI-driven interactive storytelling.

Petra's Villa is an adults-only (18+) platform. We employ multi-layered automated and human content moderation systems to enforce the standards described in this policy. This Content & Safety Policy applies to all user-generated content, including all text inputs, voice inputs, image inputs, and AI-generated outputs produced through our Service.

This policy supplements our Terms of Service. In the event of a conflict between this policy and the Terms of Service regarding content moderation and safety matters, this policy governs.

2. Age Requirement

Petra's Villa is strictly for adults aged 18 and over. By using the Service, you confirm that you are at least 18 years of age.

• Users must not misrepresent their age to access the Service
• Accounts determined to belong to individuals under 18 will be immediately and permanently terminated
• No content involving minors in sexual or exploitative contexts will be tolerated under any circumstances
• We may require you to verify your date of birth before accessing mature content
• Users who have not verified that they are 18 or older are automatically placed in PG-13 mode: all AI-generated content is restricted to exclude explicit sexual content, graphic violence, gore, drug use, and heavy profanity
• Age verification status is shared across all services operated by Old Greg’s Tavern, LLC (Petra’s Villa, Old Greg’s Tavern, and Spindle)
• Providing a false date of birth to access age-restricted content is a violation of this policy and our Terms of Service

3. Permitted Content

Petra's Villa is a permissive, adult-oriented creative platform. The AI narrator will engage with mature themes when initiated by adult players involving adult characters. The following types of content are permitted:

• Dramatic conflict and violence — Confrontations, fights, and other conflict arising in fictional storytelling
• Dark and mature themes — Death, betrayal, moral ambiguity, political intrigue, horror elements, and other mature narrative content
• Romance and explicit sexual content between adult characters — Including erotic roleplay, provided all depicted characters are adults
• Profanity and adult language
• Substance use in narrative context — Alcohol, drugs, and other substances as part of storytelling

4. Prohibited Content

The following content is prohibited and will result in content removal and potential account action, up to and including permanent termination:

• Rape, incest, or any depiction of non-consensual sexual acts or sexual violence
• Content designed to harass, threaten, or dox real individuals
• Spam, scams, or commercial solicitation through gameplay
• Attempting to manipulate, circumvent, or “jailbreak” our AI content safety systems
• Content that glorifies, encourages, or provides instructions for self-harm or suicide
• Content that seeks to extract real-world instructions for creating weapons, explosives, drugs, or other dangerous materials under the guise of in-game activity
• Sharing personal information of real people without their consent
• Creating characters that depict identifiable real people without their verifiable consent (see our Terms of Service)
• Content that violates any applicable law

5. Zero-Tolerance Violations

The following categories result in immediate, permanent account termination with no possibility of appeal. All content is preserved as evidence. Where applicable, violations are reported to the National Center for Missing & Exploited Children (NCMEC) and/or appropriate law enforcement authorities.

5b. Self-Harm, Suicide, and Crisis Situations

Petra’s Villa is a fictional roleplay game. It is not a mental-health service, and the AI narrator is not a therapist, counselor, or crisis professional. We have implemented the following safety protocols to handle interactions that reference real-world self-harm or suicidal ideation:

• Two-tier detection on every user message. Every message you send to any AI surface (the season narrator and individual character chats) passes through a two-tier classifier before reaching the story model. Tier 1 is a regex prefilter that catches unambiguous phrasings of suicidal ideation, self-harm, and crisis. Tier 2 is an AI classifier tuned to detect paraphrased and indirect expressions of real-world distress (for example, “I’m tired of being here” or farewell language). The classifier is intentionally biased toward over-detection: it would rather flag a non-crisis message than miss a real one.
• Interception and crisis-card surface. When the classifier flags a message, the story model is not called. Instead, the AI breaks character and surfaces a crisis resource card routed to the user’s region (such as 988 in the United States and Canada, Samaritans in the United Kingdom and Republic of Ireland, Lifeline in Australia, Lifeline Aotearoa in New Zealand, plus a global findahelpline.com link as a fallback for any other region).
• Privacy-protective logging. We log a SHA-256 hash of the flagged message (never the raw text), message length, surface, classifier tier, and country code, for the purpose of annual reporting under California SB 243 and continuous improvement of the classifier. We do not retain the verbatim text of flagged messages.
• Glorification prohibited. Content that glorifies, encourages, romanticizes, or provides instructions for self-harm or suicide is prohibited (see Section 4) and may result in account action.
• Limits. These systems are not infallible and operate on a best-effort basis. If you or someone you know is in crisis, please contact a qualified human professional or local emergency services. See our Safety Resources page for a list of crisis lines by region.

This Section is published in part to satisfy the safety-protocol publication requirements of applicable laws including the California Companion Chatbot Law (SB 243) and the New York AI Companion Models statute.

6. How We Detect Violations

We employ a multi-layered, real-time content moderation system that scans all user inputs and AI-generated outputs:

• Automated keyword and pattern matching for known prohibited terms and phrases
• Contextual proximity analysis to detect harmful content that may not use explicit terms but combines concerning elements
• AI-powered classification that understands narrative context, distinguishing between legitimate roleplay content involving adult characters and prohibited material
• Human review — All flagged content is reviewed by our Trust & Safety team
• Permanent violation logging — All detected violations are permanently recorded, including content, timestamps, and user identifiers

Our systems are designed to minimize false positives while maintaining zero tolerance for the most serious violations. We continuously update and improve our detection capabilities.

7. Enforcement Actions

Depending on the nature and severity of the violation, we may take one or more of the following actions:

• Warning — For first-time minor policy violations, a warning may be issued
• Content removal — Violating content is blocked or removed immediately upon detection
• Temporary suspension — For repeated violations of non-zero-tolerance policies
• Permanent ban — For zero-tolerance violations (Section 5), repeated violations, or any violation of this Content & Safety Policy or our Terms of Service that we determine warrants permanent removal. Permanent bans are lifetime bans from all services operated by Old Greg's Tavern, LLC. We reserve the right to permanently ban any user at our sole discretion.
• Law enforcement referral — For content involving CSAM, real-world threats, or other illegal activity

All enforcement decisions are logged permanently. Ban status is enforced across all services operated by Old Greg's Tavern, LLC, including Petra's Villa, Old Greg's Tavern, and Spindle.

8. Appeals Process

Users who believe their account was actioned in error may submit an appeal:

• Email innkeeper@oldgregstavern.com with the subject line “Content Policy Appeal”
• Include your account email address and a description of the situation
• Appeals are reviewed by a member of our Trust & Safety team within 5 business days

9. Reporting Concerns

If you encounter content you believe violates this policy, please contact us:

• Email: innkeeper@oldgregstavern.com
• Include as much detail as possible: account name, approximate time, and a description of the content
• All reports are reviewed confidentially
• We do not retaliate against users who report concerns in good faith

10. Policy Updates

This Content & Safety Policy may be updated periodically. Material changes will be communicated via email or in-app notification. Continued use of the Service after changes constitutes acceptance of the updated policy.

Contact

Old Greg's Tavern, LLC (owner and operator of Petra's Villa)
Email: innkeeper@oldgregstavern.com

Registered Address:
Registered Agents Inc.
30 N Gould St Ste R
Sheridan, WY 82801