Why does image-to-video reject my photo?
On Viralance an uploaded photo is refused for one of 4 reasons: the file is not a JPG, PNG or WebP; it shows a recognisable public figure; it shows a real person on a face-animation flow (lip-sync, motion transfer, DreamActor) without the consent confirmation; or it contains sexual or minor-related content. The screen runs before any credits are spent, so a refusal costs 0 credits, and if the model provider rejects the image later in the job the credits are refunded automatically. A blurry or low-resolution photo is not rejected, but it produces a weak video: use at least 1080 by 1080 pixels with a clearly visible subject.
The concrete numbers
- Accepted formats: JPG, PNG and WebP. Recommended size: at least 1080 by 1080 pixels, ideally 1920 by 1080 or larger, with good lighting and the subject clearly visible.
- Public figures are always refused, in the photo and in the prompt: generating a recognisable real person who is a public figure is blocked to prevent impersonation and deceptive media.
- An ordinary real person (you, a client, a model you hired) is allowed. On the face-animation flows, lip-sync, motion transfer and DreamActor, you confirm that you have the person’s consent before the job runs.
- The text prompt is screened too: naming a public figure, asking for a face swap or impersonation, or sexual and minor-related content stops the job before generation.
- A refusal spends nothing. If the model itself rejects the image or prompt after the job starts (a provider-side content filter), the credits for that job are refunded to your balance automatically.
- Image-to-video runs on VEO 3.1 Fast (8 credits for 6 seconds at 720p), Kling 3.0 Standard (8 credits for 5 seconds), Seedance 2.0 and 2.5, Wan 3.0 and more, with the cost shown before each run.
Common follow-ups
Can I animate a photo of myself?
Yes. A real, non-public person is allowed. On lip-sync, motion transfer and DreamActor you tick a consent confirmation first; on plain image-to-video no confirmation is needed.
Can I animate a photo of a celebrity or politician?
No. Recognisable public figures are refused in both the photo and the prompt, on every model, to prevent deepfakes and impersonation. Use a fictional or clearly original character instead.
The photo was accepted but the video looks nothing like it. Why?
Usually resolution or framing: small, blurry or cluttered photos give the model little to hold on to. Upload the largest version you have, keep the subject centred, and describe the motion in the prompt (what moves, how the camera moves) rather than re-describing the scene.
The job failed halfway. Did I lose the credits?
No. Failed generations, including provider-side content rejections, are refunded automatically. The refund appears in your credit history.
Are rejected uploads reviewed by a person?
The screen is automatic: a format check, a deterministic text screen and a vision classifier for identity. Blocks are logged for policy auditing; nobody looks at your photos to approve a job.
Pay-per-video AI generation. 20+ models, one-time credits from $9, credits never expire.
Try image to videoRelated questions
- Is there an AI video generator without a subscription?
- How do you get the lowest cost per AI video?
- Which AI video generator gives free credits without a card?
- Do AI video generators refund failed generations?
- Where can I use Kling 3.0 online without a subscription?
- How can I use VEO 3.1 online and pay per video?