This policy is part of our Terms of Service. It covers what you may put into Increation (prompts and reference images, your "Inputs") and what you may make with it (your "Outputs"). The rules are ours, and they also carry through the rules of the AI model providers we route to, which every user has to follow.
1. Don't create or upload#
Anything involving minors in a sexual or exploitative way. No sexual, suggestive or abusive content involving anyone under 18, real or fictional, under any framing. Don't upload images or voice recordings of real minors as Inputs. We will report child sexual abuse material to the authorities.
Abuse of real people's likeness.
- No sexual or intimate imagery of a real person, and no "nudifying" images of real people.
- No depicting a real person without their consent in a way that could deceive, harass, defame or embarrass them.
- If you upload a photo of an identifiable person, you need their permission to use it this way.
Deception.
- Don't present an Output as made by a person.
- Don't present an Output as a real photograph or recording of a real event when it isn't.
- Don't create content meant to mislead people about elections, public health, emergencies or public figures.
- Don't impersonate real people or organisations, and don't make a voice that imitates a real person's.
- Don't make audio that passes for a real person speaking, or for a recording of a real event.
Illegal or harmful content.
- No content that promotes or provides instructions for violence, terrorism, self-harm or serious crime.
- No hateful content that attacks people because of a protected characteristic.
- No harassment, threats or bullying.
- No extreme graphic violence or gore created to shock.
Other people's rights.
- Don't upload Inputs you don't have the right to use.
- Only include trademarks, logos and brand assets you own or are licensed to use. By including one (your own, or a client's you make work for), you confirm you have that right.
- Don't use the Service to copy someone else's copyrighted work or characters, to make counterfeits or packaging that passes goods off as another company's, or to show a real person or company endorsing something they didn't.
- Don't ask for music in the style of a named artist, or copying a named song, and don't put lyrics you don't own into a music prompt.
Explicit sexual content. No explicit sexual content of anyone, including fictional adults. Sexual content involving minors is never allowed: we have zero tolerance for it and we report it to the authorities.
2. Keep the provenance#
Some models embed information that identifies an Output as AI-generated, such as C2PA Content Credentials (OpenAI, Black Forest Labs, fal), SynthID (Google) or the inaudible watermark in ElevenLabs audio. Don't remove, alter or hide it, and don't remove visible watermarks.
3. Don't misuse Outputs or the Service#
- Don't use Outputs or the Service to train, fine-tune or build AI models.
- Don't scrape the Service, automate access outside the studio, or try to get around rate limits, credit charges or the access controls.
- Don't resell access to the Service or to the models through it.
- Don't probe, attack or disrupt the Service or our providers' systems.
4. Provider policies#
Every generation is also bound by the usage policy of the provider whose model made it: OpenAI, Google, Black Forest Labs, xAI, Recraft, Kling (Kuaishou), Alibaba Cloud, fal, Lightricks, ByteDance, ElevenLabs, Stability AI, MiniMax, PixVerse, VEED, Topaz Labs, Ideogram, HeyGen, Luma, Creatify and Sync Labs, and Anthropic for Chat and prompt enhancement (links in Terms §7). Where a provider's policy is stricter than this page, the provider's policy applies to that model.
5. How we enforce this#
- When you talk to Creatorb, what you say is turned into text, and that text is checked the same way as a Chat message before Creatorb plans, edits or spends anything.
- Before a generation runs, an automated moderation model (OpenAI's moderation service) checks the prompt, any negative prompt, and every reference image and video frame. A Chat message's text and images get the same check before the assistant replies, and so does anything you publish to the Gallery. If the check finds sexual content involving minors, explicit sexual content, threats against groups of people, instructions for self-harm, graphic gore, or help with weapons or violent crime, the request doesn't run. You see a short message saying why, and no credits are charged. Cinematic or artistic violence below the gore line is allowed. If the moderation service is briefly unavailable, ordinary requests still run, and the provider checks below still apply. Requests that look like they involve minors in a sexual way are refused until the check is back. Publishing to the Gallery waits until the check is back.
- Finished images and video clips get the same check after they're made. One that fails it is hidden behind a "Hidden by content policy" notice: it can't be viewed, downloaded, shared or reused, and we keep a record of the check. Its credits come back automatically: the whole run when everything it made is hidden, or the hidden part's share when only some of it is. This refund covers up to 3 hidden results per account in any 24 hours. After that, further hidden results stay charged and we flag the account for review.
- Chat has a second check: before the assistant plans anything, a separate AI model (Anthropic's Claude) reviews each message's text and images against this policy. If that check can't answer, the assistant doesn't make a plan.
- We keep a record of each check: the result, category scores, and a fingerprint of the prompt. For refused requests we also keep the prompt text, so we can review mistakes. We never keep a copy of the images that were checked.
- Model providers screen prompts and Outputs on their own and may refuse a generation. Refused or failed generations don't cost credits.
- We may remove content, limit features, or suspend or close accounts that break this policy, and report illegal content to the authorities.
- Apart from the automated checks above, we don't review your private studio as a matter of routine.
6. Reporting#
To report content on Increation that breaks this policy or infringes your rights, email legal@increation.ai. For copyright complaints, follow Terms §9.