How we protect your child

Alaric, the tutor, guides with questions and hints: it never hands over the direct answer to homework or an exercise. That is its default behavior, not an option you have to turn on.

When your child asks for the answer and Alaric chooses to guide instead of giving it, a visible badge appears in the conversation: "Guiding you to solve it yourself". You see it too when you read the conversation from your panel.

From your panel you can read the full conversation of every session, exactly as it appeared on your child's screen. Not a summary: everything.

There is no hidden surveillance: the study screen tells the student that their family can view the conversations.

Each session also generates a short summary written for you, which never quotes your child verbatim.

Every week you receive: what they studied, how long and on which days, where they got distracted and how many times they got back to focus on their own, and what to do the following week.

You can turn these emails off anytime in Settings.

Short sessions by default, with suggested breaks. If your child drifts away and comes back, the app rewards the return instead of punishing it: switching tabs or pausing never costs points or breaks their streak.

No message in the app uses guilt. We never mention diagnoses or tell them to "concentrate".

Every message your child writes to the tutor goes through an automatic filter (the OpenAI moderation service) that looks for signs of self-harm, sexual content involving minors and threats. If it detects something, Alaric answers with support instead of continuing the lesson, you receive an e-mail immediately with an excerpt of what they wrote and a script on how to bring it up, and the student knows at that very moment that their family will get the alert. These alerts cannot be turned off.

What the filter does not guarantee: it can be wrong in both directions, and if the service does not respond at that moment, the message goes through unreviewed so the study is not cut short (the failure is logged on our side and, while it lasts, the app generates no images). Use it as a starting point for a conversation, not as complete surveillance nor a verdict.

The tutor is forbidden from giving medical or psychological advice. FocusTutor is a study assistant and does not replace professional guidance.

The app illustrates some explanations and quests with AI-generated images. Your child never writes direct instructions to generate them: our server builds them with fixed, child-appropriate content rules.

Every generated image also goes through an independent safety filter before it is shown or stored. If the filter rejects it — or is unavailable to review it — the image is discarded and the lesson continues in text.

Generated images are private: they only exist inside your child's account and are never public.

Their hero's portrait is a generated image too, but it starts from a drawing your child builds by picking pieces from a fixed catalog (silhouette, skin, hair, outfit): the AI only changes the pose. Neither the hero's name nor any text they type reaches that generation, the portrait goes through the same safety filter, and it carries a visible "AI" badge. If you turn generated images off, the hero shows exactly as built, with no AI.

And if you prefer none at all, you can turn AI-generated images off in Settings: the app then uses only real educational photographs and diagrams. Your child cannot turn them back on; only a caregiver can.

Alaric uses your child's interests (dinosaurs, soccer, Minecraft…) to pick examples. Because that text reaches the tutor, every new interest is checked before it is saved, just like the names of subjects, exams and rewards: no hidden instructions for the AI, no inappropriate content and no personal data.

For interests, the check follows your child's age, not the age of whoever types them. Under 13, or if you didn't record their age, shooting or combat games, horror and gambling are rejected. From 13 to 17, teen-rated games such as Fortnite or Valorant are accepted, but not 18+ games such as GTA or Call of Duty.

If something doesn't pass, it isn't saved and they are asked to try another one. If the check is unavailable at that moment, it isn't saved either: we would rather wait than let something through unchecked.

Voice dictation is off until you turn it on, per child, from Settings, and we ask you to confirm before it switches on. It only works in content subjects (science, math…): in Spanish, English, French and German they always type, because there writing is the skill.

What they dictate lands as text in the same box and your child reviews it before sending: the microphone never sends on its own. The audio never goes through FocusTutor AI: we do not receive it, store it or use it to identify anyone. It is processed by the maker of their browser (Chrome: Google; Edge: Microsoft; Safari: Apple) under that company's own privacy policy, which we recommend you read; if you would rather avoid it, leave dictation off. Only the final text reaches the tutor, as if typed, and it is marked as “dictated” in the transcript you see.

In language subjects the voice has another use, also off until you turn it on and confirm: “Read it out loud”. After writing and checking an answer, or when Alaric proposes a phrase in the chat, your child can read it out loud and see which words were understood. It is not a grade. We only keep how many words were understood out of how many; never the audio or what they said. Recognition is done by the same browser service.

Your child knows it too: the first time they use the microphone we explain, in plain words, that the browser turns their voice into text and that only the text reaches Alaric.

Materials and conversations are kept in private storage. We do not sell data or train our own models with it.

Your child's first name and surnames are never sent to the AI service: Alaric knows them by their hero's name. Their real name only appears in what you see, and we add it ourselves.

There are two things we cannot control for you. What your child types in the chat reaches the tutor exactly as typed: if they write their name, their school or other people's details, that is sent too. And your child chooses the hero's name: if they use their real name, that is the name Alaric knows. From your panel you can see and change the hero's name (the pencil next to it) and read every conversation. We recommend talking with your child so they use a made-up name and do not share personal details in the chat.

The providers we use to run the service are identified in the Privacy Policy.

You can delete study history from Settings and request full account deletion anytime.

Write to support@focustutor.ai. The same father who built the app answers.