What belongs here. What doesn't. How we decide. Written in plain language because legal translations die in translation.
Every version is kept. Enforcement uses the policy version in force at the time of the action — never retroactive. See changelog.
Canon is collective. No single person — voter, admin, employee, or owner — can bend the canon of a city against the collective vote. Every piece of moderation below serves that principle first.
Removal beats silence. If we remove something, we say we did and we say why. The public audit trail opens with the moderation backend; silent shadowbans are forbidden.
Policy version matters. We enforce the version of this document that was live at the moment of the action. We don't punish you retroactively for rules that didn't exist.
Not every market can support the same content rules. The tier model below is the policy framework for the canon pipeline; live tier pages open once moderation tooling ships.
Canon publishes on schedule. Standard moderation. Public flagging. 24-hour canon cycle.
Launch framework · exact public tier list to publish with moderation tooling
Election window, social unrest, or content escalation. Canon still publishes, but with human sign-off before release. 48-hour cycle if needed.
Framework state · not a live country count yet.
Significant abuse cluster or legal constraint. Canon publishes on delay (up to 7 days) with two-admin sign-off. Users can still contribute; canon is paused.
Framework state · applied only after moderation backend is live.
No canon generated. Existing canon archives remain readable. Typically: active war, sanctions, or government takedown demand we cannot legally safely contest.
Framework state · no public frozen-country list is live today.
Living political figures: never appear in canon, micro-authoring, or user generations. Not as seeds, not as references.
Living public figures (non-political): no realistic depictions. Allowed: clearly stylized references (caricature, named archetype) that can't be confused with the real person.
Deceased figures: allowed with respect and context. Not in violent, sexual, or commercial-endorsement contexts.
Ordinary people: never by name. Ambient crowds are fine. Identifiable individuals require consent.
Public landmarks are fair game. Private property appears as fictionalised unless the owner has published it (a restaurant with a public storefront, a named neighbourhood in a novel). Trademarks: visible if they would be visible in a photograph of that place. Logos as focal subject: no.
This is where most platforms get lazy. We don't.
Islamic contexts: no figurative depiction of the Prophet, the four rightly-guided caliphs, or Qur'anic text used decoratively. Architectural and ambient Islamic imagery is welcomed.
Christian contexts: standard depictions of Christ, saints, and liturgical imagery allowed. No nudity of religious figures, no sexual content, no mockery of core rites.
Jewish contexts: Torah text not decorative. Holocaust imagery never stylised or used for effect.
Hindu, Buddhist, Sikh, Indigenous traditions: sacred figures depicted respectfully in context. No casual or commercial use. Regional tradition advisors adjudicate edge cases.
We defer to each faith's own standards, not to a universal "fairness" rule that treats all faiths identically. Treating a Jain saint the same as a cartoon character is not respect — it's laziness. Regional advisors per tradition will be published once each scope is appointed. See advisor status →
Personal image generation already uses curator-editable prompt templates. Canonical image templates will be published for audit once the public canon-render pipeline ships.
Browse the template library → Curator-maintained prompt templates.
For now, urgent flags go to safety@playdistance.com. The in-product flagging system below is the launch framework for the moderation backend.
Thresholds:
Every removal, hide, or account action will have an appeal path. Until the in-product appeal UI ships, email safety@playdistance.com; outcomes are handled manually and public logs open with the moderation backend.
Patterns matter: admins with unusually high appeal-overturn rates trigger internal review. We track our own decisions with the same scrutiny we ask of users.