PERDON VIDEO GUIDE
Which prompt makes a Perdon reveal video?
Tell the model which photo defines the subject and real ending, describe the fantasy opening, then specify the order in which each fantasy detail disappears. These are the scene instructions used in our video tool.
Perdon AI ·
MAKE MY VIDEO ↗The Paris suite
Paris suite & evening glamour → Your everyday mirror selfie
Treat reference image 1 (@image1) as the real-life ending and the source of the subject's identity. Create one continuous, natural phone-camera shot; do not make a slideshow or split screen. Start in the imagined scene, not in the reference setting. Preserve the same recognizable subject throughout. Use the selected output aspect ratio and duration. The subject is taking a mirror selfie. OPENING: an ornate gold-framed mirror in an elegant Paris hotel room, balcony doors overlooking the Eiffel Tower, a black evening outfit, pearl accessories, carefully styled hair and white roses on marble. REVEAL IN THIS ORDER: replace the view and suite with the actual surroundings from the reference; simplify the mirror to the photographed mirror; restore the photographed clothes and remove the fantasy accessories; return the hair to its everyday appearance. ENDING: the same subject takes the ordinary selfie in the uploaded setting, with the reference expression. Sound direction: gentle Parisian café-style instrumental, a brief record-stop sound at the first reveal, then quiet room ambience. Spend roughly the first quarter of the clip establishing the fantasy, the middle half revealing the changes in order, and the last quarter holding on the everyday ending. Make each change a distinct, quick visual wipe while keeping the camera movement continuous. Match the final pose, setting, clothing and expression to the uploaded photo as closely as possible. Do not add written captions, logos or watermarks; leave captions for the social editor. Keep the scene fictional and non-sexual.
The yacht life
Sunset yacht & holiday confidence → Your actual back-to-earth moment
Treat reference image 1 (@image1) as the real-life ending and the source of the subject's identity. Create one continuous, natural phone-camera shot; do not make a slideshow or split screen. Start in the imagined scene, not in the reference setting. Preserve the same recognizable subject throughout. Use the selected output aspect ratio and duration. OPENING: the subject relaxes on the deck of an unbranded white yacht, turquoise sea sparkling in warm sunset light, an open-collar linen shirt and a gold wristwatch, looking effortlessly confident. REVEAL IN THIS ORDER: transform the water, sky and boat into the ordinary location from the reference; remove the wristwatch; return the linen outfit to the exact everyday clothes; restore the photographed complexion and any glasses. ENDING: settle into the pose and framing of the uploaded photo, with a dry, tired expression only if it fits that reference. Sound direction: light summery instrumental, a short scratch-stop as the fantasy breaks, then the sound of the real setting. Spend roughly the first quarter of the clip establishing the fantasy, the middle half revealing the changes in order, and the last quarter holding on the everyday ending. Make each change a distinct, quick visual wipe while keeping the camera movement continuous. Match the final pose, setting, clothing and expression to the uploaded photo as closely as possible. Do not add written captions, logos or watermarks; leave captions for the social editor. Keep the scene fictional and non-sexual.
The champion pet
Grand champion, cape & trophy → Your pet, right back at home
Treat reference image 1 (@image1) as the real-life ending and the source of the subject's identity. Create one continuous, natural phone-camera shot; do not make a slideshow or split screen. Start in the imagined scene, not in the reference setting. Preserve the same recognizable subject throughout. Use the selected output aspect ratio and duration. Use the pet in the reference, retaining its breed, coat pattern, eye colour, face and ears, with believable animal movement. OPENING: the pet sits proudly on a velvet display stand in a chandelier-lit ceremonial ballroom, wearing a small comfortable cape, with an unbranded golden trophy and winner's ribbon beside it. REVEAL IN THIS ORDER: return the ballroom to the pet's photographed room; remove the trophy and ribbon; restore the photographed sweater or harness, or the natural coat if no garment is present; turn the display stand into the original seat or floor. ENDING: the pet sits in the uploaded setting and makes a small curious head tilt if the pose permits. Sound direction: a celebratory orchestral flourish interrupted by a quick record-stop, followed by ordinary indoor ambience. Spend roughly the first quarter of the clip establishing the fantasy, the middle half revealing the changes in order, and the last quarter holding on the everyday ending. Make each change a distinct, quick visual wipe while keeping the camera movement continuous. Match the final pose, setting, clothing and expression to the uploaded photo as closely as possible. Do not add written captions, logos or watermarks; leave captions for the social editor. Keep the scene fictional and non-sexual.
Settings and credits
The default is 12 seconds, 720p and 9:16 portrait. Advanced settings offer 4–15 seconds, 480p or 720p, and 9:16, 16:9 or 1:1. Any supported combination uses one video credit. APIMODELS provides Seedance 2.0 Mini.
See one-time video pricing · Existing photo credits? Open the photo maker.
Keep the same request when a task is delayed
Check again retrieves an existing task. It does not generate another video. A confirmed failure returns the reserved credit; an uncertain submission is held for checking, with undelivered requests released after 24 hours. Download completed results before the 24-hour retention period ends.
Before you share
AI may change faces, pet markings and other details. Exact identity or final-frame matching is not guaranteed. Use consenting subjects, keep the fiction clear, and add only audio you have permission to use. Privacy notice · Acceptable use
FAQ
Can I change these prompts?
Yes. Selecting a story loads its prompt in the video maker, where you can edit it before generating. A new generation uses one video credit.