The content standards, moderation, reporting and enforcement rules that govern AI generation through reAPI.
2026/08/10
Last updated: August 10, 2026
reAPI provides API access to AI models developed by third parties. While those providers apply their own safeguards, every user remains subject to this Acceptable Use Policy (this "Policy"). By using the Services you agree to comply with it. This Policy forms part of the Terms of Service, and violating it can result in action against you or your account, up to termination and legal referral.
You are fully responsible for your own use of the Services and for that of your end users, including every prompt, reference file and output.
This Policy applies to all text, images, video, audio, prompts, reference material and outputs generated, submitted, stored or distributed through the Services.
Don't use the Services, or any input or output, to generate, promote or further any illegal, unlawful or indecent activity or content, including:
Don't use the Services to engage in, promote, incite or facilitate uncivil content, including:
Don't use the Services for fraudulent, deceptive, defamatory or otherwise unethical activity, including:
Don't use the Services in a way that:
Don't use the Services in any way or for any purpose that:
Don't use the Services to engage in, facilitate or generate criminal activity, including the creation of illegal or regulated goods, or any other criminal conduct.
Don't use the Services to compromise personal, sensitive or confidential information, or to threaten the integrity of the Services, including:
The following result in immediate termination, and we reserve the right to report them to the competent authorities. No account tier, contract or configuration grants an exception:
reAPI is an aggregation layer over third-party models, so moderation operates on two levels. We describe both accurately rather than claiming more than we perform.
Provider-side safety systems. Every request is executed by the model developer's own infrastructure, which applies that developer's content classifiers to the prompt and its output review to the result. These are the safety systems named on each model's page and in its provider documentation, for example ByteDance's content checks on the Seedance family and Google's on the Veo family. They run on every route without exception and are the primary technical control. A request refused on content grounds fails as a terminal error and the reserved credits are refunded.
Platform-side controls. Requests submitted through the browser playground run with provider safety checking enabled and cannot disable it. Some models expose a moderation setting to direct API callers; using it changes which route executes the request, and the upstream model's own checks still apply. It does not exempt anyone from this Policy, and everything prohibited above remains prohibited on every route.
Reference material. On some models, reference images, video and audio are submitted for automated review before generation runs. Where that review applies, rejected material fails with an error and a full refund.
Rate limits. API keys carry configurable request limits, which bound the volume any single credential can generate.
Report-triggered review. Automated checks run before and during generation, so they cannot catch what only a person looking at the finished result would recognize — an unconsented likeness, a defamatory claim, output presented elsewhere as authentic footage. That gap is covered by §11: anything reported to us is reviewed by a person against this Policy. Every request is retained with its task identifier, which is what lets us locate the material a report refers to, confirm which account produced it, and act on the account rather than only on the single output.
We do not claim to operate our own content classifier, a dedicated trust and safety team, or hash-matching partnerships. Review of reports is carried out by reAPI staff. Where the other capabilities exist elsewhere in the industry, we will say so plainly if and when we adopt them.
Where to report. Email [email protected], or use our contact page, which reaches the same inbox. This is a monitored address, not an automated endpoint — every report is read and decided by a person. Start the subject line with REPORT so it is routed ahead of ordinary support mail.
What to include. Give us as much of the following as you have:
You may use the same address to request removal of a stored output that depicts you, or that uses your voice, likeness or copyrighted work without your permission.
How we handle a report. Receipt is acknowledged, the report is triaged into one of the severity bands below, the material is reviewed against this Policy by a person, a decision is taken under §12, the action is carried out, and the reporter is told the outcome. Where a report is upheld we disable access to the stored output and act against the originating account — not only against the single item reported.
| Severity | Scope | Initial response | Resolution |
|---|---|---|---|
| Critical | Anything in §9, or a credible threat to life | Within 24 hours | Within 72 hours |
| High | Severe violation of §2 through §8 | 3 business days | 7 business days |
| Standard | All other reports under this Policy | 5 business days | 15 business days |
Reports concerning child safety or credible threats to life are triaged ahead of everything else and are not held in a queue. If a case needs longer than the window above, we tell the reporter and say why.
How we treat reporters. Your identity is not disclosed to the reported account. A report is assessed on the material itself, so reporting a competitor does not by itself cause action against them. Reports submitted in bad faith, and repeated submission of reports already found unsubstantiated, are recorded and may themselves be treated as a violation of this Policy.
| Severity | Action |
|---|---|
| First minor violation | Written warning and removal of the offending output |
| Repeated or deliberate violation | Suspension of API keys and forfeiture of remaining credits |
| Severe violation | Immediate account termination, forfeiture of remaining credits, and record retention where required |
| Zero tolerance (§9) | Immediate termination and referral to the competent authorities |
We may suspend an account while an investigation is in progress.
If you believe an action was taken in error, email [email protected] within 30 days of the notice, with APPEAL at the start of the subject line, your account email and the relevant task identifier. The material is examined again against this Policy rather than against the reasoning of the earlier decision, and we respond with the outcome. Actions taken under §9 are not subject to appeal.
An API accepting an upload is not a determination that you hold the rights to it.
We may update this Policy as regulation and model capabilities change. Material changes are reflected in the "Last updated" date above.
Reports of a violation (§11), appeals (§13) and questions about this Policy all reach us at [email protected], which is also the address behind our contact page. Prefix the subject line with REPORT or APPEAL so the message is routed accordingly.