Safety filters run on every generation. If your prompt or the image you uploaded trips one, the generation is stopped before anything is created. Results can also be reviewed after creation, so content that slips through a filter can still be removed later.
Common reasons
The filters look for the categories listed in what you can and cannot create: sexual content, nudity, graphic violence, anything involving minors, hate speech and harassment, impersonation, and content resembling public figures. They check your uploaded images as well as your prompt.
Filters work on signals, not intent, so an innocuous prompt can occasionally land near a prohibited category and get stopped. Rewording the prompt to be more specific about what you actually want usually clears it.
What happens to your credits
What comes back depends on why the generation didn't complete:
Technical failures, 100% back. If the model errored, the job was interrupted, or no output was produced, all credits for that generation return to your balance automatically, usually within minutes. If they didn't, email support@magicshot.ai within 7 days with your account email and the date.
Moderation blocks, 50% back. If the block happened because our system detected an NSFW prompt or an NSFW input image, 50% of the credits return automatically. The other half covers the cost of processing the blocked attempt and is not refunded. A moderation block is not a technical failure.
False positives, everything back. Filters act on signals, not intent, and occasionally stop an innocent prompt. If you appeal and human review confirms your generation didn't violate the content policy, the withheld 50% is restored in full.
The full rules live in the Refund Policy.
If you think the block was wrong
Ask for a human review. See appealing a moderation decision. If the review agrees with you, your withheld credits come back.
