Safety
Protection is built into how tinybuilder works, not added as a warning.
Last updated 2026-06-09 · draft — under review before public launch.
Nothing goes public by accident
Before a child’s app appears in the store it passes an automated content scan (source + a visual check), a grown-up’s approval, and a moderation review. A trusted track record can streamline the review, but a parent’s consent is always required.
Guardrails are built in
The AI never collects or repeats a child’s personal information, never produces harmful content, and gently steers back to building. Published apps run fully sandboxed with no network access and no external links.
No dark patterns
No streaks, no follower counts, no notifications designed to pull a child back, no pressure of any kind. A child comes when they want to make something and leaves when they’re done.
Anyone can report
See something that doesn’t belong? Every published app has a report control — no account needed. Reported apps are pulled for re-review. You can also email hello@tinybuilder.agent.