Safety & young users

LevelUp is for people aged 13 and over. If the law where you live sets a higher minimum age for a service like LevelUp, you must meet that age.

To keep what we hold about you to a minimum, LevelUp doesn't ask for your email address, phone number or date of birth. You use the app without signing in: the app gives your account a random ID instead, and you choose a nickname, which doesn't have to be your real name. We recommend that it isn't, especially if you're under 18.

LevelUp is not directed to children under 13, including under the US Children's Online Privacy Protection Act (COPPA), and we don't knowingly collect personal data from them. If we learn that an account belongs to a child under 13, we suspend it and then delete the account and its data. Suspending it also marks the app installations it was used on, and the marker stays after the deletion, so a new account made on that device is suspended too (see Suspended devices).

If you are a parent or guardian and think your child under 13 is using LevelUp, email [email protected] with the subject "Child under 13". Because accounts have no email or password, please include the Support ID from the app on your child's device (Profile › Settings › Help & Contact › Support IDProfile›Settings›Help & Contact›Support ID; tap it to share or copy it). Without it we may not be able to find the account. If you can't get it, tell us what you know, such as the child's nickname. We will tell you if we can't find the account, and we may ask you for more details before we act. You can also delete the account straight away on that device: Profile › Settings › Privacy & Data › Delete accountProfile›Settings›Privacy & Data›Delete account. If you ask us for a copy of the child's data instead, we will also ask you to show that you are the child's parent or guardian.

LevelUp's AI plans are written by Claude, an AI made by Anthropic, not by a person. You don't chat with the AI. When you ask for a plan, the app sends Claude the details of your quest, Claude writes a week-by-week plan, and the plan comes back to you to review and edit. The weekly rebuild works the same way: it uses how your week went to rewrite the next week's tasks.

• Nothing is sent for a plan until you allow it. Before the first AI plan or weekly rebuild, the app shows what will be sent and who receives it: "Nothing about your quest is sent for a plan until you allow it." Nothing about your quest goes to Anthropic for an AI plan until you tap Allow. "Not now" sends nothing to Anthropic. If you choose "Not now" for the free AI plan during setup, that quest tracks progress without XP; the app says so and suggests adding a ready-made starter quest that has rewards.
• You can withdraw at any time in Profile › Settings › Privacy & DataProfile›Settings›Privacy & Data. Withdrawing stops sending your quest details to Anthropic for AI plans and weekly rebuilds, and stops a plan that is still being written. Plans you already have stay in your quests.
• Never your name, email or account details. Anthropic receives the quest details needed to write the plan, but not your name, email, account details or any ID. Don't put personal details in what you write for the AI.
• No training. Anthropic doesn't use this data to train its models.
• Community text is checked separately. Your nickname, the quests you share and your comments are checked automatically by Anthropic's Claude before they're published, whether or not AI plans are on. Turning AI plans off doesn't stop this; Privacy & Data says: "Your nickname, shared quests and comments are still checked by Anthropic's Claude, as the Community Guidelines say." Only the text being checked is sent (for a nickname check, the nickname itself), never your account ID. See Automatic checks of community text.

Our Privacy Policy lists exactly what is sent, how long it is kept and where it is processed, and our Terms of Use cover how AI plans may be used.

In AI plans

• Safety rules for the AI. The instructions LevelUp gives the AI tell it not to write plans with extreme dieting or rapid weight loss, long fasts, purging or diet pills, dangerous stunts or training through pain, or illegal, sexual or self-harm content, and not to diagnose, recommend medication or give medical advice. When someone mentions a health condition or injury, the AI is told to keep the plan conservative, and to decline when no safe version exists.
• Checks before anything reaches the AI. LevelUp checks each plan request for wording about self-harm, eating disorders and prolonged fasting. If it finds any, it declines the request without sending it to the AI.
• Declined plans point to help. When the AI won't write a plan, the app suggests trying a safe, healthy version of the goal or writing the tasks yourself. When the request touches on self-harm or eating, it also encourages the person to talk to a doctor or a local support line. Our Community Guidelines list crisis lines by country.
• Report this plan. Every plan the AI wrote has a "Report this plan" link under it: in New Quest, in the plan preview during setup and on the rebuilt-week screens. Pick what's wrong (Dangerous advice, Self-harm or suicide, Sexual content, Violence or illegal activity, Hate speech, or Something else: "Wrong, misleading or inappropriate in another way"), add details if you like (optional, up to 500 characters) and send it. The report goes to our moderators with a copy of the plan, including any extra details you gave the AI, and helps us improve the AI's safety rules. The copy is kept with the report until you delete your account. See AI plans.

In the community

• Reporting. Anyone can report a quest, a comment or a nickname from the "…" menu in the app, and anyone can report content by email through our content report page, with or without an account.
• Automatic hiding. Some reports hide content at once while we review it; the rules are in What we do.
• Review by a person. A moderator reviews every report. We aim to review each one within 24 hours, and child-safety reports are treated as urgent.
• Enforcement. We remove content that breaks our rules, remove nicknames and suspend accounts. When we suspend an account, the app installations it was used on are marked, so a new account made on them is suspended too. For child-safety and intimate-image reports we keep a secure copy of the evidence. See What we do.
• Less to misuse. LevelUp has no private messaging, and other users can't see your profile photo. The community shares text only: quests, comments and nicknames.
• Blocking. You can block another user with Block [name] in the "…" menu on their quest or comment. Neither of you then sees the other's quests or comments or can comment on the other's, and they aren't notified. You can unblock them in Profile › Settings › Blocked UsersProfile›Settings›Blocked Users. See Blocking someone.
• Automatic checks before publishing. Shared quests, comments, nicknames and rebuilt weeks of shared quests are checked automatically by Anthropic's Claude before they're published. Text that doesn't pass isn't published. If the check fails because of an error on our side rather than a short outage, the text is published without it, so reports and moderators remain the main safeguard. See Automatic checks before publishing.

For your data

• No ads. LevelUp doesn't collect the advertising ID and doesn't track you across other companies' apps or websites.
• Usage analytics can be turned off with "Share usage analytics" in Profile › Settings › Privacy & DataProfile›Settings›Privacy & Data.
• Marketing notifications ("Updates & offers") are off unless you turn them on, during setup or later. You can turn them off any time in Profile › Settings › NotificationsProfile›Settings›Notifications.
• You can delete your account at any time in the app: Profile › Settings › Privacy & Data › Delete accountProfile›Settings›Privacy & Data›Delete account.

• Talk to a doctor first before you start a new exercise or diet program, especially if you are under 18, pregnant, or have an injury or a health condition. Stop and get help if something hurts.
• Keep health details out. Don't write health or medical details, or your name or contact details, in the "Anything the AI should know?" box or in your quest titles and tasks.
• Think before you share. If you share a quest made with AI, read it first: everyone in the community will see it.

Every screen that shows a plan the AI wrote (the New Quest plan review, the plan preview during setup and both rebuilt-week screens) says "AI-generated, may contain mistakes; check before relying on it." Fitness, sports and cooking plans, and plans without a category, also say "Not medical advice: check with a doctor first." The "Report this plan" link sits under these lines.

If something in LevelUp sexualizes, exploits or puts a minor at risk:

1. In the app: open the quest, tap the "…" menu on the quest or comment, choose Report quest or Report comment (or Report followed by the person's name, for a nickname), pick Child safety and tap Send report. It alerts our moderators as urgent. If your account is at least 24 hours old and has finished setup, the content is usually also hidden at once while we review it (a few limits against misuse apply).
2. By email: write to [email protected] with the subject "Urgent: child safety". Describe where the content is. Do not attach, copy or forward the material itself.

Our Child safety standards explain how we handle these reports and list organizations that can help.

We are committed to complying with the child-safety and privacy laws that apply to LevelUp in the countries where it is offered, including Türkiye's Personal Data Protection Law No. 6698 (KVKK) and, for users in the EU and EEA, the EU General Data Protection Regulation (GDPR). How we process personal data under KVKK is explained in the KVKK Aydınlatma Metni (in Turkish). We follow the Usage Policy of our AI provider, Anthropic. Where its guidelines for organizations serving users under 18 apply to LevelUp, in line with them we tell you that plans are written by AI, not a person, we use safety rules and checks for AI plans and community text, and we publish this statement.

When a safety measure is added or changes, we update this page, its version and the change log at the top.

Questions about safety: [email protected]. Who we are: Support.