> ## Documentation Index
> Fetch the complete documentation index at: https://safety.flamy.net/llms.txt
> Use this file to discover all available pages before exploring further.

# Principles

> The values that anchor Flamify policies and moderation decisions.

# Principles

> The values that anchor our policies and moderation decisions.

## 1. Protect people first

Content that depicts, sexualises, or targets minors is never allowed.
Self-harm, real-world crime facilitation, and exploitation are treated as
hard stops — not edge cases.

## 2. Adults may create adult fiction

Flamify is an adult AI companion product. Consensual adult erotic roleplay
between fictional adult characters is allowed on the main product surfaces.
Policy is about harm and legality, not about erasing adult fiction.

## 3. Separate creative surfaces from distribution surfaces

What is appropriate in a private adult chat is not automatically appropriate
on an app-store or PG-13 brand front. [Safe filters](/moderation/safe-filters)
exist so distribution surfaces can stay SFW while the adult product remains
available where intended.

## 4. Automate volume, escalate judgement

Keyword and LLM classifiers handle volume at speed. Human reviewers and the
trust team handle ambiguity, appeals, and patterns that classifiers alone
cannot resolve.

## 5. Policy is independent of growth metrics

Moderation decisions are made independent of revenue and acquisition goals.
The trust team owns the policy; product owns the product.

## 6. Be honest about limits

Classifiers are imperfect. Fail-open and fail-closed paths are documented.
We improve rules from false positives, false negatives, and reports — we do
not pretend the system is infallible.