Acceptable Use Policy

Version 2026-09-06

This policy is part of the Terms of Service. It says what you may and may not make with Bluffy, and it is honest about what is enforced by software and what is a rule you are trusted to follow.

Questions and appeals: hello@bluffy.ai.

1. The posture

Bluffy is a professional film tool. Mature, dark, violent and sexual content is legitimate filmmaking and is not filtered. We do not screen your work for taste, we do not apply a general content filter to your prompts, and we do not review your outputs.

Only two things are enforced by software, and each is narrow on purpose:

one absolute prohibition, checked on every render request (section 2);
the rules of a provider that holds a durable asset for you, checked at the moment that asset is saved (section 3).

Everything else in this policy is the law, or a provider's terms, restated so that you know where the edges are.

2. The absolute prohibition

You may not attempt to generate sexual content involving minors. There is no artistic exception, no fictional exception, and no exception for any framing.

This is a serious crime, and it would also get the provider accounts that every Bluffy user shares terminated at once.

Bluffy runs a narrow, high-precision check for this one category on the prompt of every render request. It is tuned to leave legitimate dark and mature work alone, and it is the only content filter in the product. Three blocked attempts within 24 hours pause generation on the account until the window clears, and each block is recorded.

If a legitimate prompt is blocked, that is a bug in our list and we want to know: hello@bluffy.ai.

3. Real people, and assets saved at a provider

You may reference real people in your prompts. Sometimes a production genuinely needs it, and Bluffy does not stand between you and your script. We do not scan prompts for names and we do not scan outputs for faces.

The line is drawn where something is saved at a provider under our account. Two features do that:

Registering a Virtual Character. Saving a character as a Virtual Character registers its identity images with the model provider (BytePlus ModelArk today; other providers may follow), which holds them as a standing asset under Staiple's account and Staiple's word. That provider's rules bind there, and they do not permit a registered character to depict a real, identifiable person without that person's consent. So: you may not register a character built to depict a real, identifiable person — living or dead, famous or not — without that person's documented consent, or the consent of their estate where the right survives.
Registration runs automated checks: whether the image is a photoreal human; where the prompts that created it exist, whether they asked for a real person; and, for a photoreal human, an automated comparison of the image itself against a celebrity face database. Any of them can refuse, and a refusal is recorded. A character uploaded without the prompts that made it is not refused for that; the image itself is what is judged.
Working around these checks — laundering an image through another tool, splitting a likeness across steps — is a breach of this policy whether or not it succeeds.
Cloning a voice. A cloned voice is stored as a voice asset by the provider, under our account, and that provider's rules require the speaker's consent. You may not clone a real person's voice without their consent. A voice can only be cloned from a sample already in your own Bluffy Library — a take you rendered or a clip you uploaded under your own attestation — never from an outside link.

Outside those two features, likeness is your creative and legal call, and the law (section 5) still applies to what you make.

Withdrawn voices. We withdraw preset voices from the casting pool when they are named after real performers or copyrighted characters. A withdrawn voice stops working immediately, including in projects that already used it.

4. Rights in what you upload

When you upload an image, a video or an audio recording, you confirm that:

you own it or are licensed to use it this way, and
every identifiable person in it has consented to this use, and
it may lawfully be processed by the providers listed in the Privacy Policy, outside Canada.

Bluffy asks you to attest to this when you import media and when you build a character from an upload. The attestation is the basis on which the upload is permitted. Attesting falsely is a breach of this policy, and the consequences of it are yours.

Do not upload material you obtained without permission, however easy it was to obtain.

5. The law still applies

Bluffy does not check for any of the following. You are trusted not to do them, and doing them is a breach of this policy:

Break the law where you are or where we are, or help someone else to. That includes making non-consensual intimate imagery of a real person, material built to defraud or to harass, and material that defames.
Create material designed to deceive about a real person or event and pass it off as genuine — fabricated evidence, a real person appearing to say something they did not, forged documents or identity material.
Attempt to circumvent Bluffy's safety checks, spend limits or account boundaries; access another user's content or assets; or probe the service for vulnerabilities without written permission. If you find one, we want to hear about it: hello@bluffy.ai.
Resell access, share your account, or run Bluffy as a service for others.
Use automation to burn provider capacity at no cost to you — for example by repeatedly cancelling renders after they have started.

6. Providers have their own rules

FAL and the model vendors behind it, BytePlus, and the vendors behind the Vercel AI Gateway each have their own acceptable-use terms, and each can refuse work independently of us. A refusal you did not expect may be theirs rather than ours. Their rules apply to you in addition to this policy, and they are sometimes stricter.

7. What happens on a breach

Depending on what happened, we may refuse a generation, pause generation on the account, suspend the account, retire an asset saved at a provider, or close the account. Where the law requires it, we will report to the authorities.

We keep records of refusals as evidence that the checks ran.

If you think a refusal was wrong, tell us. A false positive that blocks legitimate work is a failure of ours, and we would rather fix the rule than leave you stuck: hello@bluffy.ai.

8. Changes

We will post changes here with a new version date. Where a change is material, we will ask you to accept it before you continue.

Staiple Technologies Inc. · West Vancouver, British Columbia, Canada · hello@bluffy.ai