child safety standards

Last updated: 21 July 2026.

This is our published standard against child sexual abuse and exploitation (CSAE). It is a requirement of the Google Play and Apple App Store developer policies, and it is a document we would publish anyway.

Our standard, in one line

oomfy ai has zero tolerance for child sexual abuse and exploitation, in any form, real or fictional, drawn or described, and we ban and report without warning.

We are 18+ only, and there is no minor experience at all

This is the foundation of everything else on this page, so it is worth being precise about it.

oomfy ai is not an app with an adult mode. It is not an app with a filter you can turn up. There is no under-18 version of this product to get into. Age is attested at signup and is a gate on every route in the system: no chat endpoint will open without it.

That single architectural decision is what lets us treat every CSAE signal as an unambiguous emergency rather than a moderation judgement call, and it is why our answer here is a ban rather than a queue.

What is prohibited

All of the following result in an immediate, permanent ban:

  • Child sexual abuse material (CSAM) of any kind.
  • Any sexualisation of a minor, including fictional, drawn, animated or textual depictions. A character card that declares an adult age but presents as a child is prohibited. A character described in a way that sexualises childhood is prohibited. "It's fiction" is not a defence, and neither is a number in a metadata field.
  • Grooming, solicitation, or any attempt to make contact with a minor.
  • Sexualised roleplay in which any participant is presented as a minor.
  • Content that normalises, promotes, or provides instruction for the sexual abuse of children.
  • Trading, requesting, or linking to any of the above, anywhere on the service — in a message, a character, a persona, a lorebook, a name, a tag, an image, or a bio.

There is no strike system for any of this. There is no appeal on the merits.

How we prevent it

  1. The age gate. 18+, attested, enforced server-side on every authenticated route. Not a checkbox on a splash screen.
  2. Character moderation fails closed. Every character, and every edit to a character's prompt-bearing text, is moderated before it can be published. If the moderation verdict cannot be parsed, the character is rejected — never silently approved. A system that fails open on this is a system that will eventually publish the worst thing anyone ever sent it.
  3. Image generation is scanned before delivery. Generated image bytes are not served to anyone until they have been through a safety scan. The scan fails open to a human review queue — a provider error or an unreadable image goes to PENDING for a person and is never auto-approved.
  4. Every user turn is classified before a reply is generated.
  5. Prompt-injection resistance. A character definition cannot instruct the model to ignore our safety scaffold — creator text is spliced around our system prompt at a defined token, never concatenated blindly over it.
  6. Reporting is one long-press away, on every message and every character, and it reaches a human.

How we respond

When we identify CSAE content or conduct:

  1. The content is removed immediately.
  2. The account is permanently banned, along with any other account we can associate with it.
  3. We preserve the evidence as the law requires.
  4. We report to NCMEC (the National Center for Missing & Exploited Children) and to any other authority the relevant jurisdiction requires, including India's law-enforcement agencies under the IT Act and POCSO.
  5. We cooperate fully with any resulting investigation.

We do this whether or not the content was ever shown to another person, and whether or not it was generated privately.

Reporting to us

[email protected]

This mailbox is monitored, it is prioritised above every other queue we have, and it reaches a human being. If you are reporting CSAE, please include as much detail as you safely can — the character, the account, a timestamp. You do not need an account to write to us, and you will not be asked to explain yourself.

If a child is in immediate danger, contact your local emergency services first. We are a company. They are the people who can actually go there.

Our point of contact

The named point of contact for child-safety matters, for both the Google Play and Apple App Store CSAE standards, is reachable at [email protected].

Where else to report