Safety

What we do to keep this appropriate for children, and what to do when something is not.

Written into the instructions

Every story plan is generated under a fixed set of rules that the model receives with each request. Among them: no unresolved darkness, no threatening figures, no injury, and no peril the text has not already made safe. The reading level is matched to the age band you choose, and the story is required to open on a character doing something rather than on a threat.

You see the plan before any picture exists

This is the most useful safety feature in the product, and it is structural rather than technical. Planning happens first, before any illustration. You read the whole story — every page, every line — and can rewrite anything before a single illustration is generated. Nothing is drawn without your explicit approval.

Image filtering

The image model applies its own safety filtering, which we leave enabled. Prompts also carry explicit instructions against text, logos and frightening imagery. Filtering is not perfect; if you get an illustration that is wrong for a children's book, repaint the page and let us know.

Privacy by default

Stories are private when you make them. Publishing is always a deliberate action, produces an unlisted link rather than a public listing, and can be undone. Nothing you create is added to the public library.

The library

The stories in our library were produced by the same pipeline under the same rules and reviewed before publication. They are free to read and require no account.

Report content

If you find anything in the library that is not right for children, or if a generation produced something it should not have, tell us at [email protected] with the story link or id. We review every report and remove library content that does not belong.