Acceptable Use Policy

Who and what this covers

This policy applies to every user, signed in or not, and to every input and output: story prompts and ideas, character descriptions, uploaded reference photos, generated illustrations, generated text, book titles, and anything you publish to the public gallery or share by link. You are responsible for what you submit, including that you have the right to upload any photo you use as a reference.

Prohibited content

You may not use Doodara to request, generate, upload, publish, or share any of the following. Each is prohibited on its own:

• Pornography and NSFW content — sexually explicit or sexually suggestive imagery or text, nudity presented sexually, and fetish content.
• Violence and gore — realistic or graphic depictions of injury, torture, mutilation, killing, self-harm, or suicide, and content that encourages any of these.
• Hate speech — content that attacks, demeans, dehumanises, or incites hatred or violence against a person or group on the basis of race, ethnicity, national origin, religion, caste, sex, gender identity, sexual orientation, disability, or serious disease, including hateful symbols.
• Child sexual abuse material (CSAM) — any sexual or sexualised depiction of a minor, in image or text, without exception, including stylised, cartoon, or AI-generated depictions. This is the category we enforce most aggressively and the only one we report to the authorities.
• Deepfakes and impersonation of real people — generating a recognisable real person, living or dead, without their consent; using someone else's photograph as a reference without permission; and presenting generated content as a real person's words, likeness, or endorsement.
• Content that infringes copyright or trademarks — reproducing copyrighted characters, artwork, or text you do not have the right to use, and using protected brands, logos, or trade dress.

Zero tolerance

The following end an account immediately and permanently on the first occurrence, with no warning, no appeal to a warning stage, and no refund of any fees paid:

• Child sexual abuse material, or any attempt to sexualise a minor.
• Content that promotes, glorifies, or facilitates terrorism, violent extremism, or mass violence.
• Instructions for weapons capable of mass harm.
• Non-consensual sexual deepfakes targeting a real person.
• Repeated, deliberate attempts to defeat our safety checks, including jailbreak prompts and obfuscated spellings intended to slip past screening.

How we screen prompts

Screening happens before an image or story is generated, not after it is published, and it runs in two layers because each covers the other's blind spot. The first layer is a deterministic check against a multilingual list of explicit terms, applied to your story prompt, your character description, and any look hint you supply; it runs before any model sees the text, costs nothing, and cannot be talked around. It also applies a stricter combination rule that rejects sexualised phrasing about minors even when neither half would be a problem alone. The second layer is a classification step performed by the model that reads your prompt, which is required to return a content flag and catches phrasing no word list can anticipate. A prompt rejected at either layer is not generated, and the attempt — the prompt, the character description, the reference image URL, and the language it was written in — is recorded so a human can review it. The list targets explicit and abusive content only: it is deliberately not a filter for difficult subjects, because picture books legitimately deal with death, illness, divorce, fear, and bodies, and blocking those would fail the families we are for. We also review reports and publicly shared books, and we may re-review anything at any time.

What happens when a rule is broken

We apply the least severe measure that fits, except in the zero-tolerance categories above, where the maximum applies immediately:

• Rejection — the prompt is refused and nothing is generated. This is the normal outcome of a screening hit and is recorded as a strike against your account.
• Warning — for a first minor breach, we tell you what was wrong and what will happen if it continues.
• Content removal — a book, illustration, or gallery entry that breaches this policy is unpublished or deleted, and any public link stops working.
• Account suspension — for repeated breaches. Three recorded violations suspend an account automatically. A suspended account cannot generate anything.
• Permanent termination — for zero-tolerance breaches or repeat offences after a suspension. Fees are not refunded for an account terminated under this policy.
• Referral to the authorities — content involving child sexual abuse material is reported to the appropriate authorities, and we preserve the associated records.

Reporting a violation

Email report@doodara.com with a link to the book, image, or gallery entry, and a sentence about what is wrong with it. You do not need an account to report something, and you do not need to be the person affected. We aim to acknowledge every report within one business day, and to act on child-safety reports immediately, ahead of everything else. Reporter identity is kept confidential and is never shared with the account reported. Every report is reviewed independently of who sent it. Reports we find to be deliberately false or made to harass another user are themselves recorded and actioned. Report:

• Content that breaches any rule in this policy.
• Anything involving child safety or terrorism — these are handled first, at any hour.
• Content that uses your likeness, your child's likeness, or your copyrighted work without permission.
• An account you believe is systematically abusing the service.

If we get it wrong

Automated screening makes mistakes, and a false rejection on an innocent story is a real cost to a real family. If your prompt was refused and you believe it should not have been, or your account was suspended and you think that was an error, email hello@doodara.com with the prompt you used. A person reads it. If we were wrong we will say so, remove the strike, and fix the rule that caused it. Terminations for zero-tolerance breaches are not appealable, other than to establish that the account was compromised or that the wrong account was actioned.

Contact

Reports of violating content: report@doodara.com. Anything else about this policy, including an appeal: hello@doodara.com. Both are monitored by a person.