Notice on the use of Artificial Intelligence (AI) features

Pufffi · pufffi.hu / pufffi.com · v1.0 · Effective: 8 July 2026

Some Pufffi features use artificial intelligence (“AI”). This notice explains what the AI is for in Pufffi, its limitations, how we moderate content, how we protect children, and what happens to your data. This notice forms part of Pufffi's Terms of Service and Privacy Policy and should be read together with them.

1. Which AI features does Pufffi have?

• Task breakdown (text): splits a bigger task into small, doable steps.
• Task breakdown (photo): from an uploaded image (e.g. a messy room, an appliance panel) it suggests steps or guidance (PRO feature).
• “Break into steps”: splits an existing task into even smaller micro-steps.
• “Explain it”: explains a given topic in plain language.

Pufffi's AI is not an open, free-form chatbot: it can only be used for the targeted tasks above.

2. Limitations of the AI — the AI can be wrong

The AI's answers do not constitute expert, medical, psychological, legal, financial or other professional advice, and are not suitable for making a diagnosis or replacing a professional. For health, legal or other professional matters, consult a qualified expert. When using home appliances, the manufacturer's original manual prevails.

3. Content moderation

Every AI response goes through built-in moderation, in both adult and child accounts, to keep use safe and on-purpose:

• The task breakdown only helps with household, errands and similar everyday tasks. It will not write finished schoolwork, essays, translations or code (for studying it only provides the steps of the process).
• “Explain it” classifies each request by content and refuses adult or harmful topics: 18+/explicit content, drugs, encouraging alcohol or tobacco use, self-harm/suicide, violence, making weapons or explosives, illegal activity. Questions about the body, reproduction and puberty are explained at a factual, biological level.

Moderation is automated and, like any automated system, is not perfect: it may occasionally allow or block something in error.

4. Protecting children

Pufffi is also made for family use, so children's safety is a priority:

• Topics that aren't child-friendly (see section 3) are blocked; the child then receives no answer, but a caring message pointing them to a trusted adult.
• For self-harm and abuse, instead of an answer we show the official, free helpline details.
• If a request about a sensitive topic comes from a child profile, we send an immediate email alert to the parent account's email address. Exception: for domestic violence we do not send an alert, because the account owner may be involved; in that case the child is shown a neutral helpline.
• Kids mode can also be locked with a PIN.

These measures help but do not replace parental supervision. The AI is not infallible, so we recommend that a parent supervises a child's use of the AI.

5. Data processing related to the AI

To operate the AI features, Pufffi sends the content you provide (the text you type, or the image you upload for photo breakdown) to the AI provider for processing. We do not send any account identifier (your name, email address, user ID) to the AI provider, and because the request originates from our server, it does not receive your IP address either. The text or image may, however, contain personal data if you include it — so please do not enter or photograph unnecessary personal or special-category data.

• AI provider (data processor): OpenAI, L.L.C. (United States; EU representative: OpenAI Ireland Ltd). On Pufffi's behalf, via the API, the provider processes the submitted content and returns the output (the steps/explanation).
• Model training: per its policy, OpenAI does not use data submitted via its business (API) service to train its models; it may retain the data for a limited period for abuse monitoring and then delete it.
• Photos: images uploaded for photo breakdown are not stored by Pufffi and are not used to build any gallery; after processing they are not kept on our server.
• Transfer to a third country (USA): the transfer takes place with appropriate safeguards (the European Commission's Standard Contractual Clauses, SCCs, and — where applicable — the EU–US Data Privacy Framework).
• Please do not share sensitive data: avoid putting unnecessary personal or special-category data into the text or the uploaded image.

The legal basis is the provision of the service (performance of a contract) and — for moderation and the child-safety alert — the legitimate interest of the Controller and data subjects in safety. Detailed data-processing rules, your rights and the other data processors are set out in the Privacy Policy.

6. Fair use, allowance and limitation

AI features depend on your plan and come with a monthly allowance (credits) to prevent abuse and overload. The Controller reserves the right to change the availability, allowance or operation of the AI features, and to restrict access in case of abusive use.

7. Liability

The AI features are provided “as is”. To the extent permitted by applicable law, the Controller accepts no liability for the accuracy, completeness or fitness for a particular purpose of AI-generated content, nor for decisions made based on it. You use the AI's answers at your own risk; please always verify important information as set out in section 2.

8. Age and parental responsibility

Creating and managing a Pufffi account is the task of an adult (parent/guardian). Child profiles are used under the parent account with the parent's supervision. The parent is responsible for supervising their child and for the kids-mode settings (e.g. PIN).

9. Changes to this notice and contact

The Controller may amend this notice unilaterally; the current version is available at pufffi.com. Questions: hello@pufffi.com.