← Wizard Maker

Content Policy

Last updated: 13 September 2026

Apps built here are shared with other people and run on their phones. These rules cover what you may build, what the AI will refuse, and what happens when something crosses the line. They apply to prompts, to generated apps, to the data inside them, and to anything you send through the platform.

Not allowed

Child safety

Anything that sexualises a minor, or that endangers, exploits or grooms a child. There is no framing under which this is acceptable. Accounts are terminated on the first offence and reports are referred to the relevant authorities.

Sexual content

Pornography, sexually explicit imagery, and apps whose purpose is sexual roleplay or explicit generation.

Allowed: dating apps with ordinary profiles, relationship and intimacy guidance, sex education, and health apps that deal with bodies frankly.

Violence and extremism

Instructions for weapons, explosives or attacks; recruitment for or glorification of terrorism or violent extremism; gratuitously graphic or gory material; threats against a real person or group.

Allowed: news and history covering violent events, hunting and shooting-sport logistics, self-defence and personal-safety apps, fiction and games with non-gratuitous conflict.

Hate and harassment

Content attacking people because of race, ethnicity, national origin, religion, disability, sex, gender identity, sexual orientation, age or immigration status; and tools built to harass, bully, dox or brigade a person or group.

Allowed: apps that study, report on, moderate or counter hate and harassment.

Self-harm

Encouraging or instructing suicide, self-harm or disordered eating, and content that glorifies them.

Allowed: mood and recovery journals, crisis-resource directories, eating-disorder recovery support, and clinical or educational material.

Illegal goods and services

Marketplaces or sourcing tools for illegal drugs, unlicensed weapons, counterfeit documents or currency, stolen data or credentials, or protected wildlife.

Allowed: medication reminders and dosage trackers, licensed-firearm and hunting-permit records, harm-reduction and substance-recovery apps.

Gambling

Wagering real money or anything of real-world value. This is a licensed, geographically restricted activity that Wizard Maker cannot support, so it is refused rather than gated.

Allowed: card and casino games with play money and no purchases, and odds or probability calculators.

Deception and fraud

Phishing pages, scams, fake reviews, pyramid and Ponzi schemes, and impersonation of a real person, business, bank or government service.

Allowed: parody clearly labelled as parody, fraud-awareness training, and security-education material.

Privacy and surveillance

Covertly tracking, recording or monitoring a person without their knowledge — including partner-tracking apps; bulk scraping or harvesting of personal data; and publishing someone's private information.

Allowed: location sharing where every person involved knows and agrees, family-safety apps with visible consent, and asset or fleet tracking you own.

Malicious behaviour

Malware, exploit tooling, credential theft, network scanning, denial-of-service, cryptocurrency mining without consent, and anything designed to attack a system or evade its defences.

Misleading advice

Presenting AI output as authoritative medical, legal or financial advice. Generated apps can be wrong, and an app that hides that is a hazard.

Allowed: symptom and habit trackers, budget tools, and reference apps that are clear about what they are.

How we enforce this

There are four layers, and none of them is only a keyword list:

1. Before generation. Prompts are screened for unambiguous violations. This layer is deliberately narrow so that ordinary app ideas are never caught by it.
2. During generation. The model is instructed to judge the purpose of a request, not its vocabulary, and to refuse anything on this page. This is the layer that does most of the work, because intent is something a model can weigh and a pattern cannot.
3. After generation. Finished apps are scanned before they are stored, so a refusal the model failed to make does not become a shareable app.
4. Reports from people. Every app and every AI response can be reported from inside the Wizard Maker app. Reports go to a human queue.

What happens when something is reported

We review every report. Reports involving child safety, credible threats, or self-harm are looked at first. Outcomes are:

• No action — nothing here broke the rules.
• Removal — the app is taken down, for everyone it was shared with.
• Suspension or termination — for serious or repeated breaches.

Child-safety violations end the account immediately and are referred onward. For everything else, we normally tell you what was removed and why, and you can appeal to appeals@wizardmaker.com.

How to report something

In the Wizard Maker app, open the app in question and choose ⋮ → Report this app, or ⋮ → Report AI response for something the AI produced. Pick a reason, add detail if you want to, and you can block the person who shared it at the same time.

Blocking is symmetric and silent: a blocked account cannot share an app with you and you cannot be added to theirs, and they are not told. Manage your blocks in Settings → Blocked accounts.

If you cannot use the in-app flow, email abuse@wizardmaker.com.