VRBO stills are the listing; generating a new room is a policy problem
Animate the approved listing photo. Do not invent amenities.
Animate the approved listing photo. Do not invent amenities.
The stills already on the VRBO listing are the unit a guest booked against. Motion on those stills is honest. A text-to-video kitchen with a breakfast bar you do not have is a different property. That is not a style upgrade. That is a listing problem.
AI image to video treats the upload as frame one. What plays at 0:00 is the file you sent. Camera and light are the only things the prompt is allowed to add.
The listing still is the advertised unit
If you would not put the still in slot one, do not buy seconds of it. A dark phone snap of the wrong wall is not a brief. Approve the photo: even light, the actual furniture, the actual view out the actual window. Then animate that file.
Image-to-video keeps composition, trim, and layout. Beds, fireplaces, and the number of sinks survive because they were already pixels. Text-to-video has to invent the room. It will invent a better sofa, a second island, a balcony that faces water. Guests do not book a vibe. They book the photos.
Lock the still before you spend on motion is the credit argument. Video meters are a terrible identity debugger. They are also a terrible listing-accuracy debugger. Fix the jpeg. Then spend the clip.
Amenities you typed are not amenities you have
A hot tub, a pool, an ocean view, a loft, a king in the second bedroom — those are listing facts. If they are not in the approved stills, do not prompt them in. The model will draw them. The guest will look for them. Generating a new room is how you advertise amenities you did not photograph.
The Airbnb hosts playbook is the same job on the same kind of property: a walkthrough of the unit you actually operate, not a generated lobby. VRBO stills are not a different contract because the booking site is different. The stills are still the unit.
Prompt only motion: a slow push through the living room you photographed, curtains that were already in the frame, late light that was already on the floor. Do not prompt "add the spa" or "make the yard look larger." Empty is honest.
Reference-to-video is the other miss. Image-to-video vs reference is the split: I2V is this frame; R2V recasts the subject in a new scene. Recasting a rental in a kitchen you never shot is how a cabin becomes a different cabin. Do not restage the listing.
Walkthroughs start from rooms you own
A useful host clip is still a walkthrough of the real layout. Under a minute, room to room, from stills you already shot for the listing. Image-to-video can add a push or a drift on those stills. It cannot add a bathroom.
Do not invent occupancy. A generated family at the table is a household you do not have a grant for, and it hides the furniture a booker is trying to see. House rules, parking, and Wi-Fi belong as type you wrote after picture lock, not as chrome a video model typesets.
The test: would a guest who pauses the clip see the same room they will walk into?
Yes → animate the listing photo.
No → you generated a new room. Throw it away. Shoot the real one.
FAQ
Can I text-to-video a "better" living room if I describe the listing carefully?
You can. You will recast furniture, windows, and the view. The expensive failure is an amenity that is not on site. I2V the approved listing photo so those are not sampled again.
Is a generated pool or sunset view just marketing?
No. It is an amenity claim. Guests book against the photos. Animate the view you photographed. Do not invent a different one.
Does image-to-video replace the listing stills?
No. Slot one is still the photograph. The clip is supporting motion of that photograph. Do not swap in a generated still as the main image.
Should I generate guests using the rooms?
No. The unit, empty or with a crew hand you filmed, is the product. Invented households are people you do not have a grant for, and they cover the layout a booker is scanning for.