Content Safety Policy

Ouchy is an adult service with hard limits. These restrictions apply to AI-generated and computer-generated content exactly as they apply to any other content.

Absolutely prohibited

• Minors, or persons represented as under 18, in any sexual context — zero tolerance
• Sexual exploitation of minors — zero tolerance
• Non-consensual sexual activity, rape or sexual assault — zero tolerance
• Sexual activity involving a person who is unconscious, incapacitated, drugged, intoxicated or otherwise unable to consent
• Incestuous sexual activity
• Sexual activity involving animals or non-human creatures
• Extreme sexual violence, graphic torture, mutilation or sexualised violence
• Trafficking or exploitation
• Sexual content involving real persons without appropriate rights or consent
• Impersonation intended to facilitate fraud, abuse, harassment or exploitation
• Hate-based or discriminatory content
• Illegal activity, or content designed to facilitate criminal activity

How moderation works

Safety is enforced on the server, independently of the model's system prompt. Every user message entering the AI pipeline passes through authentication, entitlement, rate limiting, safety classification, the context engine, the model, output safety classification, response delivery and logging.

Voice is not an exception: speech-to-text transcripts are subject to the same safety system as text chat, before and after model generation.

Moderation policies are versioned. We store moderation result metadata rather than unnecessary raw sensitive content.

What happens on a violation

Ouchy may refuse, block or modify a response, end a session, restrict features, suspend the account, or report the matter to appropriate authorities where legally required.