Air fryers: the basket still cannot become a different SKU

I2V the listing photo.

An air-fryer ad dies when the basket in motion is not the basket on the PDP. Handle, window, capacity marks, the crisper plate in the box — those are the product. Text-to-video will redraw them. Image-to-video is how they stay. Animate the listing photo. Do not prompt "our 6-quart matte black basket" and hope the hinge survives.

Seedance 2.5 Image to Video is the still-to-motion row: one locked frame into a continuous clip, up to 30 seconds, up to 720p, native audio. A second image sets the closing frame when the shot has a destination — drawer closed, then open on the food. Image required. Forty-seven credits listed.

The listing photo is the brief

If you would not print the still on a box, do not buy motion. Even light, basket large in frame, label readable. That still is cheaper to argue about than a 20-second clip with the wrong handle.

Seedance I2V starts there. Camera and steam are the prompt. Identity is already in the pixels. Repeating "exact SKU" in text is how you invite a renegotiation.

Thirty seconds in one pass is why this family belongs on a PDP loop. Stitching four short T2V clips is how the window tint drifts by second twelve.

720p is the ceiling. Most feed and PDP loops live there. A 2K hero you will grade is a different generate, still from a locked still if you can.

Reference-locked ads are a different door

The AI product video generator is for shots where the object must remain yours while the scene changes: the basket on a counter you never rented. That path is reference-to-video. Two or three clean angles are the contract. Fine print is the first thing to inspect at 100%.

Do not mix contracts. I2V means this listing frame is the first frame. R2V means identity travels into a new room. Feeding a finished hero still to R2V and asking it to "just add a sizzle" will restage.

Seasonal re-skins belong on the product-video path: same references, new scene. A new SKU is a new still, not a prompt adjective. The basket cannot become a different SKU because you typed "family size."

What the prompt is allowed to do

Motion only: drawer sliding, vapour, a slow orbit. Aspects include 9:16, 1:1, 16:9. Generate vertical first if the ad is a Reel.

Native audio is on. Write the mix: drawer slide, faint sizzle, no speech — or you get a band you did not ask for. A talking presenter holding "your" fryer is a different row.

Last frame is optional. Use it when the open-basket food shot is a second approved photo. Do not attach a last frame of a different model to "show capacity." That is how SKU A closes as SKU B.

Lock the still before you spend on motion. Video meters are a terrible identity debugger.

Pause on the handle

Pause. Read the window, the handle, the plate. If any would fail a listing audit, throw the clip away. Do not "fix" it with T2V. Make a new still. If they match, caption after. The basket stayed the basket because you never asked the model to invent it.

FAQ

Can I text-to-video an air fryer if I describe the listing carefully?

You can. You will re-cast the basket every take. Labels, handles and window tint are the expensive failure. I2V the listing photo so those are not sampled again.

Is 720p enough for a PDP loop?

For feed and most product-page loops, yes. For a 2K plate headed for a grade, pick a higher-resolution generate and still start from a locked still. Resolution is delivery. Identity is I2V.

How many photos does the product-video path need?

Three or four is usually enough: front, three-quarter, any branded side. Even light matters more than volume. That is the reference-to-video door, not the single-still I2V door.

Should the food in the basket be generated?

Only if the still already has the food you will ship, or you accept that the food is atmosphere. If the recipe is the SKU (a branded liner, a specific tray) lock that in the photo. Do not let motion invent a second accessory.