AI Content Policy

• Some features run locally in your browser or on our own servers. These call no external provider and cost no credits. They are labelled as free in the interface.
• Paid features send your prompt and inputs to an AI provider. The credit cost is shown before the action runs.
• With Bring Your Own Key, the request runs under your own provider account and is billed by them.

The Sub-processors page names every provider and what each receives.

AI output is probabilistic. The same prompt can produce different results, and results can be inaccurate, biased, stereotyped, or unexpectedly similar to existing work. Output may include text artefacts, anatomical errors, or invented detail presented confidently.

As between you and Obrasken, you own your prompts and your output. Two limits are worth understanding properly rather than discovering later.

1. Provider terms may restrict how output can be used, including for commercial purposes. Those terms apply to output produced through that provider.
2. In several jurisdictions — including the United States and Canada — material generated without sufficient human authorship may not attract copyright protection. You may be free to use it and still unable to stop others from doing the same.

Output is not exclusive. Another user with a similar prompt may receive similar material.

• Do not generate a real person's likeness or voice in a way that suggests they said or did something they did not.
• Voice cloning requires the consent of the person whose voice it is.
• Commercial use of a real person's likeness generally requires their permission, separate from anything in these policies.
• Do not upload other people's personal data or images without a lawful basis for doing so.

Where it could mislead — journalism, advertising, political material, commerce, or anything presented as documentary — disclose that content is AI-generated or AI-assisted. Some jurisdictions and platforms now require this. Beyond compliance, it is simply the honest thing to do.

Requests pass through provider safety systems and our own checks. Requests that breach the Acceptable Use Policy may be blocked, and false positives happen — appeal via Support. Report harmful output that gets through so we can act on it, either from the tool or by email.

Every paid action shows its cost before running. Where a paid action fails before producing a result, we aim to leave your balance unaffected. Disliking a completed result is not a failure, because the provider cost was still incurred — see the Refund Policy.