Acceptable Use Policy
Last updated:
Short version: write your dreams, publish what you are willing to show a stranger, and do not use the generator to make something you would not want your name on. This page is the detailed version, and it applies to everything you put into Dream Memos and everything it produces for you.
1. What this policy covers
It covers three things: the text you submit (a dream, a follow-up question, a Plaza post), the content our models produce from it (readings, dream reports, answers, and generated images), and the way you use the service itself. Where this policy and the Terms of Service overlap, both apply.
2. Content that is not allowed
The following is not permitted, whether it appears in text you submit, in an image you ask us to generate, or in something you publish to the Plaza:
- Sexual content involving minors. Any sexualization of a person under 18 — text, request or image. This is reported to the relevant authority where we are legally required to do so.
- Adult sexual content. Explicit sexual descriptions or imagery, including requests to generate it.
- Violence, threats and abuse. Content that threatens a real person, celebrates or instructs in violence, or depicts serious harm to people or animals.
- Hate and harassment. Attacks on people for their race, ethnicity, religion, disability, sex, gender identity or sexual orientation; targeted abuse of an individual.
- Deepfakes and impersonation. Making it look like a real, identifiable person said or did something they did not, or pretending to be someone else.
- Other people's private information. Real names, addresses, phone numbers, ID numbers, health records or intimate images of someone who has not agreed to their publication.
- Copyright and trademark infringement. Reproducing protected characters, artwork, logos or passages you have no right to use, or asking us to recreate them.
- Illegal and dangerous activity. Instructions for making weapons, drugs or malware; how to seriously harm yourself or someone else; fraud, scams, or the sale of prohibited goods.
- Breaking the service. Bypassing rate limits, quota, content checks or the account system; scraping; automated bulk requests; reselling access; probing for vulnerabilities outside a coordinated disclosure.
3. Extra rules for generated images
Image generation has its own failure modes, so these are called out separately. Do not ask us to draw:
- an identifiable real person, a celebrity, or a private individual's face;
- identity documents, banknotes, coins, stamps, signatures or official seals;
- a well-known copyrighted character, logo or brand mascot;
- anything designed to deceive — fake screenshots, fake news images, fake evidence.
A generated image is yours to keep and to use, but that permission does not extend to content that breaks this policy: if an image is made in breach of these rules, the licence does not cover it.
4. How content is reviewed
Every dream you ask us to illustrate is checked by an automated classifier before any image is generated. The final image prompt is checked again after rewriting, and a vision model checks the generated image before delivery. Every Plaza post is checked before it becomes visible. We retain the verdict, policy and model version, and a fingerprint, rather than another copy of your private content.
Automated checks are not perfect in either direction. We also act on reports from people who see something on the Plaza, and we may review a generated image manually when a report or a pattern of abuse warrants it. If a check is unavailable, we say so rather than silently passing everything: a failed image check prevents delivery and restores the image allowance.
5. Reporting something
Every Plaza post has a report control. You can also write to us directly, at:
Please include the URL or a screenshot, and say what rule you think is broken. Reports are read by a human.
We aim to act on reports of the most serious categories (anything involving minors, credible threats of violence) within 24 hours, and on other reports within a few days. We will not tell the person you reported who reported them.
6. What happens when the rules are broken
Depending on severity we may: remove the content; restrict the feature that produced it; suspend the account; or close it. Where the law requires it, we report the content and what we know about the account to the relevant authority. Repeat breaches escalate — the first response to a borderline case is usually a warning or a removal, not a ban.
7. Appealing
If you think a decision was wrong, reply to the same address in section 5 with the content in question. A different person reviews appeals where we can; if the service is small enough that this is one person, we will say so rather than pretend otherwise.
8. Changes
We may update this policy as the product and the law change. The date at the top moves with it, and changes that materially widen what is prohibited are announced on the site rather than edited in silently.