Child Safety & Wellbeing Policy

Famara accounts are for adults. An adult who owns a Famara family may add a family member under 18 and mark them as a child ("under-18 account"). By doing so the adult confirms they are the child's parent or legal guardian, consents on the child's behalf to our Terms and Privacy Policy, and accepts responsibility for the child's use of Famara. We record the time that consent was given.

These protections are part of the Famara platform itself. They apply the same way to the model Famara chooses automatically, a model the family picks, and a family member's own connected key.

• Checked before answering. Every chat message from an under-18 account — including attached documents, checked in full — is checked by an automated safety screen before any AI model sees it. Messages longer than 12,000 characters are not sent to the AI at all. Earlier turns are taken from Famara's own saved, already-checked history, and turns the safety guard answered are never sent to an AI later. Messages about serious risk get crisis help even when the family's monthly allowance is used up.
• No photos to the AI. Under-18 accounts cannot send photos to Famara's AI, because images cannot be screened first.
• Crisis help, not an AI answer. Messages about self-harm or suicide, abuse or grooming, eating disorders, sexual content, drugs or hurting others receive a fixed, caring reply written by Famara — with Australian help lines — and no AI model is asked to answer them.
• Safe by default when checks are unavailable. If the safety screen cannot run, the child receives a short holding reply instead of an unchecked answer.
• Answers are checked too. AI replies to an under-18 account are checked before they are shown; an unsuitable reply is held back.
• Rules that cannot be switched off. Age-appropriate rules are applied by our servers to every answer. Instructions sent from the account that try to change them are dropped, and saved memories or custom instructions are not used on under-18 accounts.
• Features limited. Under-18 accounts cannot use Deep Research, model comparison, web search, the app builder, image generation, web-page reading, publishing or sharing chats publicly, chat widgets, the browser extension, memory, Community or friend features, and cannot be found or contacted by other users.
• No escaping supervision. An under-18 account cannot leave its family, create its own family, delete its own account, or change its own age setting. Only the parent can — and marking a member as an adult requires the parent to confirm they are 18 or over.

• Pause instantly — one tap on the Family page stops chat and the AI features for that child, enforced on our servers, until resumed.
• Safety alerts — when a concerning message is caught, the parent receives an in-app alert and an email with the category and a short excerpt, and can see recent alerts on the Family page.
• Remove a child from the family at any time.

If a child tells Famara that a parent, carer or family member is hurting them, Famara shows the child Kids Helpline, 000 and eSafety reporting details and records that the event happened, but does not alert the family owner and does not show them what the child wrote — because a person who may be causing harm could otherwise see it, which could increase the risk to the child. This protection then stays on for that child's account for 90 days: follow-up messages are not shown to, or alerted to, the family owner either, and the owner cannot open that child's conversations during that time. Urgent events of this kind are escalated to Famara's own safety contact instead, so that a responsible adult is still told. A temporary fault in the safety check never switches this protection on.

For each safety event we keep the category, the time, whether it was caught before or after an answer, whether the parent was alerted, and a short excerpt (withheld where the child reported harm by a family member) — as a record that the event was handled. A child who is not in a family is still protected, but there is no parent to alert. Pausing, resuming, and changing a member's age setting are recorded in the parent's notifications. See our Privacy Policy.

Automated safety screening reduces risk; it cannot remove it. It may miss something, or flag something harmless. Famara does not replace parental supervision, professional advice or emergency services. Parents remain responsible for their child's use of Famara and of their devices.

Famara is operated from Western Australia by AXD Intelligence (ABN 36 242 601 526). This policy is designed around the Online Safety Act 2021, the eSafety Commissioner's industry codes for generative-AI and chatbot services (in force from 9 March 2026), the Basic Online Safety Expectations, and safe-messaging guidance on suicide and self-harm. Concerns: legal@famara.ai. You can also report harmful online content to the eSafety Commissioner at esafety.gov.au/report.