Acceptable Use Policy

Last updated: 30 October 2026

This policy is part of the vincevangen Terms of Service. It says what you may not create with vincevangen and how you may not use the Service. It applies to prompts, input files, outputs and everything done with your account or API keys, including by AI agents acting for you. An agent acting on your instructions is you, for this policy.

Each model provider also has its own usage policy (for example Google's Generative AI Prohibited Use Policy and Black Forest Labs' Usage Policy). Those apply too; the models page links them.

1. Content you must not create or upload

We check prompts, and input images for edits and video, with an automated safety model before any paid call. These categories are blocked:

1.1 Child sexual abuse and exploitation (zero tolerance)

• Any sexual or sexualised content involving a minor or anyone who appears to be a minor, real or fictional, including drawn, animated or "aged-up" characters.
• Content that grooms, sexualises or exploits children in any way.

A request in this category is refused, the account is flagged for human review and suspended, and we report to the relevant authorities where the law requires or allows it.

1.2 Sexual content

• Sexually explicit or pornographic images, video or audio, including nudity in a sexual context. The hosted Service blocks all sexual content, not only content that is illegal.

1.3 Harmful imagery of real people

You may create imagery of real people, for example editing a photo of yourself or your team. You must not create:

• sexual, nude or intimate imagery of a real, identifiable person, even if they agree (we cannot check consent);
• imagery that puts a real person in a degrading, violent or false compromising situation, or that harasses them;
• imagery that impersonates a real person or is presented as real in order to deceive (deepfakes), such as fake endorsements, fake evidence or election disinformation;
• sexualised or otherwise harmful imagery of a real child;
• any of the above from a photo of someone.

1.4 Violent extremism

• Content that promotes, glorifies or recruits for terrorist or violent extremist groups or acts, or that depicts real attacks for that purpose.

Our safety model may also flag other categories (for example hate, self-harm, weapons or criminal instructions). Those are covered by section 2 and acted on after review.

2. Other prohibited uses

Even where our filters do not catch it, you must not use the Service to:

• break any law, or help someone else break it;
• infringe intellectual property: for example, generate counterfeit logos or trademarks, or copy a living artist's or a studio's protected work in order to pass it off;
• harass, threaten, bully, or incite violence or hatred against people based on protected characteristics;
• promote or give instructions for self-harm, suicide or eating disorders;
• create content that facilitates weapons capable of mass harm, or other serious crime;
• deceive: fake documents, IDs, receipts or evidence; impersonating a real person or organisation; election or public-health disinformation; presenting AI imagery as a real photo of an actual event;
• remove or hide watermarks or AI markers a model provider requires (for example SynthID);
• use outputs to train or improve an AI model that competes with the maker of the model that made them;
• publish AI-generated content without labelling it where the law requires a label;
• spam or mass-produce content for fraud, fake reviews or manipulation.

3. Using the API fairly

• Keep to the limits. Per-key rate limits, per-account concurrency and job limits apply; the dashboard shows the current values. Do not get around them with extra keys or accounts.
• No reselling raw access. Do not offer the Service, or a thin wrapper around it, to others as a generation API, proxy or model endpoint. Building your own product with the assets you generate is fine.
• No key sharing. Keys are for your account's own people and agents. Do not share, sell, publish or embed them in software others install. Revoke a leaked key at once.
• One account per person for the sign-up gift. No multiple accounts to collect free credits, no disposable emails to get them.
• No attacks. Do not probe, scan, overload or attempt to break into the Service, other accounts' data or our providers, and do not try to bypass moderation, tenant isolation or spending limits (including by splitting a blocked prompt across requests or encoding it).
• Security research: welcome, with permission first. Write to [email protected].

4. What happens if you break this policy

Depending on how serious it is, we may: refuse the request; withhold an output; warn you; rate-limit, suspend or close your account; revoke keys; remove credits obtained by abuse; and report to authorities. Child-safety violations always lead to suspension and review. Blocked requests cost no credits.

We record each block: time, account, key, the categories and a fingerprint (SHA-256 hash) of the prompt, not the full prompt. We keep these records for 2 years. Withheld outputs are kept for review and not delivered.

Automated moderation makes mistakes. If you think a request was wrongly blocked or your account was wrongly suspended, write to [email protected] and a person will look at it.

5. Reporting

To report content made with vincevangen that breaks this policy, or a leaked key, email [email protected]. Include the asset id (a_...) or the provenance record if you have it; every image we deliver carries one in its metadata.

6. Changes

We may update this policy as laws, our providers' policies or abuse patterns change. Material changes are announced by email and in our changelog. Changes for safety or legal reasons can apply at once; other material changes get the 30 days' notice in the Terms.