Perdon AI videos.
Your reality reveal.

It is a fantasy-to-reality reveal: a glamorous invented scene opens the video, then the setting, props and outfit change one by one until the subject is back in the everyday situation shown in a real photo.

One. Upload an everyday photo of yourself, a consenting adult or your pet. It supplies the subject and real-life ending; you do not need to create a separate fantasy picture first.

No. This workflow sends one reference photo and scene instructions to the video model. The example clips illustrate the story and are not uploaded with your generation request.

No. Select the photo, generate and download the MP4 in your browser. Add captions and publishing audio in your preferred social editor afterward.

Yes. Choose a story, add your photo, confirm permission and select Generate video. Seedance 2.0 Mini is accessed through APIMODELS.

Browsing examples, choosing templates and copying prompts are free. Generating a video requires one video credit; there is no free video generation allowance.

Video packs are $4.99, $11.99, $17.99 for 1, 3 and 5 videos respectively, in USD before applicable tax. Each video credit covers one delivered Seedance 2.0 Mini clip, 4–15 seconds at 480p or 720p. No subscription or automatic renewal.

APIMODELS is the provider and Seedance 2.0 Mini is the video model. The default is 12 seconds, 720p and 9:16 portrait. You can select 4–15 seconds, 480p or 720p, and 9:16, 16:9 or 1:1. One video credit covers any of these settings.

Yes. Open Scene instructions, edit the prompt and review it before generating. Changing a template replaces the scene instructions. The model and billable limits are controlled by the server.

Yes. The champion-pet template starts in a fictional awards setting and returns the pet to its photographed room, clothes and seat. Use a clear photo with the face, ears and coat markings visible.

The prompt asks the model to preserve the subject and return to the reference setting. AI output can vary and may change identity or details, so exact matching is not guaranteed. Review the result before sharing.

No licensed trending song or written captions are included. The prompts request model-generated background sound. Add your own permitted audio and Perdón captions in TikTok, Reels or another editor.

A confirmed failure returns the reserved video credit. If submission is uncertain, the same request stays reserved and is not automatically sent again. If no result is delivered within 24 hours, the credit returns. Check again retrieves an existing task without starting a new generation.

No. Photo and video balances are separate. Existing photo credits keep their original value and remain usable in the photo maker. Your Google account or recovery code restores both balances.

Selecting a photo keeps it in your browser. Generate sends a resized, metadata-stripped copy to our server and APIMODELS. For video references our server holds it only in memory, behind a random provider-access URL, for at most 45 minutes. It is not saved to our database or disk. See Privacy for upstream processing.

Your private generated MP4 is available for 24 hours. Download it before expiry. You can delete a finished result earlier; task metadata is kept for 30 days. Provider copies follow APIMODELS' own policy.