Updated 23 October 2026 · AI ceiling fan transition · your own clothes, in order

AI Ceiling Fan Transition With Your Own Outfits

The AI ceiling fan transition, also called the under-the-fan or fan blade outfit change, is an overhead clip of someone lying on a rug beneath a turning ceiling fan: each time a blade sweeps across the lens they reappear in a new outfit, while the face, pose and room stay put. People filmed it for real with a phone fixed to the ceiling early in 2026; the AI version spread on TikTok in late September 2026. This page makes it with clothes you actually own.

One selfie + four outfit photos · 15 s · 9:16 · from 84 credits (about $0.83) · no watermark · failed videos refunded

Your photo

One clear photo of your face, looking at the camera, no sunglasses; waist-up or full length works best. JPG, PNG or WebP, 300–6000 px per side, up to 20 MB.

Outfits

Look 1

Look 2

Look 3

Look 4

One outfit per photo, in the order they should appear: Look 1 opens the video, Look 4 closes it. You can pick several files at once. Worn photos, flat lays, hanger shots and shop product photos all work.

Model
Resolution
Sound

Trend sound lays the template’s own 15-second track under your video. Pick Silent to add your own music in TikTok or Reels. Same price.

480p · 9:16 · 15 s MP4 · four looks

This video costs 84 credits. Charged when you start, refunded automatically if it fails.

Example of this template.

What is the AI ceiling fan transition?

The camera points straight down from the ceiling. A person lies on the floor while a ceiling fan turns between them and the lens, close enough that a blade covers most of the frame for a split second. That moment of cover is the cut: once the blade has passed, the clothes are different. You will see it called the ceiling fan transition, ceiling fan outfit transition, fan transition, under the fan or fan blade outfit change.

It started as a filmed trick. Creators fixed a phone to the ceiling facing down, recorded one take per outfit in the same spot and joined the takes where a blade crossed the lens; an Instagram tutorial by @colinpomeroyyy, who credited @jeffmachad0, was circulating early in 2026. In late September 2026 the AI version took off on TikTok (The Tab wrote it up on 28 September), because a generator can make the shot from a photo, with no fan, no ladder and no outfit changes.

There is no single official sound. Duran Duran’s Girls on Film turns up often, and many posts use whatever is trending that week. This tool lays the template’s own 15-second track under your video, or leaves it silent for your own music.

This generator at a glance

QuestionAnswer
What you uploadOne photo of your face, plus four outfit photos (or none, if you pick a theme pack).
Order of the looksThe order you upload them: Look 1 opens the video, Look 4 closes it.
Length and frame15 seconds, 9:16 vertical, one person.
Resolution480p, 720p or 1080p, depending on the model.
SoundThe template’s own track, or silent. MiniMax H3 always has sound.
PriceFrom 84 credits (about $0.83) with Seedance 2.0 Mini at 480p. Own outfits and theme packs cost the same.
Watermark and failuresNo watermark on any video. A video that fails gets its credits back automatically.

Which outfit photos work

The model reads each outfit photo for the garments only: colour, cut, fabric and print. It is told to ignore the person, mannequin, hanger, background and any text in the photo. What matters is one complete outfit, clearly visible.

Photo typeWorks?Tip
You wearing it (mirror shot or a friend’s photo)Yes, the most reliableFull length, not cut off at the knees or feet, with no bag hiding the front.
Flat lay on a bed or the floorYesLay out top, bottom and shoes together and shoot from straight above in daylight.
On a hanger or railYes, for single piecesHang top and bottom together; put the shoes beside them if you want them in the look.
Shop product photo on a modelYesThe model’s face is ignored. Choose the shot that shows the whole garment from the front.
Screenshot from a shop, Pinterest or TikTokUsuallyCrop out prices, buttons and captions: text left in the image can come back as a fake print.
Several outfits in one collageNoSplit it up: one outfit per photo, one photo per look.
Close-up of a fabric or a detailNoThe model needs the whole garment to know where a detail sits. Keep close-ups for checking the result yourself.

Which look appears when

The template has four looks and three changes, each hidden behind a passing blade. Your outfit photos fill them in upload order.

SecondsOn screenComes from
0 – 5.4Look 1, while you roll off the sofa and settle on the rugOutfit photo 1
5.4 – 9.0Look 2Outfit photo 2
9.0 – 12.9Look 3Outfit photo 3
12.9 – 15.0Look 4, the shortestOutfit photo 4

Put the outfit you most want remembered first or second: those get the longest screen time. Each look replaces the whole outfit, shoes included, so a photo without shoes may get shoes the model picks. Four different photos are needed; the same photo twice is refused.

What stays and what can drift on your clothes

A video model redraws every frame, so it holds on to the overall garment better than to small detail. The table is what to expect from a reference-clip template like this one, based on how these models behave, not a lab test of this page.

DetailUsuallyWhat helps
Colour and overall cutKeptEven daylight in the outfit photo; no strong colour filters.
Plain fabrics, denim, knits, leatherKeptNothing special.
Large prints and stripesMostly kept; the pattern can shiftA sharp, front-on photo without creases across the print.
Small prints, embroidery, laceCan be simplified or redrawn720p or 1080p, and a close, high-resolution outfit photo.
Logos and text on clothesOften garbledExpect them to change; plain-front pieces are safer.
Sequins, sheer and shiny fabricsThe shine stays; the exact pattern can changeA photo where the fabric lies flat in even light.
Long drapes: dupattas, saree pallus, capes, trainsCan be shortened or rearrangedShow the drape laid out in the photo, and expect it to fall differently on the rug.
Shoes, bags and jewelleryShoes kept if shown; accessories not in the photo are not meant to be addedInclude the shoes; leave out anything you don’t want worn.

If a detail matters, for a sale or a brand post, compare the finished video with the real garment before posting and say the video is AI-generated.

Festive, wedding and cultural outfits

Many of the most-shared versions of the trend show occasion wear: four áo dài, a run of salwar suits, wedding guest looks, kimonos. They work the same way, with a few extra points.

  • Áo dài, cheongsam and other long fitted dresses: a full-length photo matters more than usual, because the side slits and the hem are part of the look.
  • Saree, lehenga and salwar kameez: photograph the outfit worn or fully laid out, dupatta or pallu included. A folded saree gives the model too little to go on.
  • Hanbok, kimono and robes with wide sleeves: the silhouette carries over well; small woven motifs can be simplified.
  • Wedding guest and bridal looks: put the main outfit as Look 1 or 2 for the longest screen time. Veils and trains may come out shorter.
  • Religious and ceremonial dress: use photos you are entitled to use, and leave out sacred items you would not want an AI to redraw.

Your face in all four looks

The template clip supplies only the room, the fan, the camera and the movement. Your face photo is the only source of who is on screen: the instructions sent to the model keep the same face, hair, skin tone and build in every frame and take nothing from the person in the template.

What can still drift: the face is small in an overhead shot, so fine features soften at 480p; hair can change shape between outfits; glasses or a beard you don’t have in the photo are not meant to appear but occasionally do. A sharp, front-facing photo in even light and 720p or 1080p help most. A generation that fails is refunded automatically; a finished video you simply don’t like is not.

How to make it in your browser

  1. Open this page on your phone or computer and sign in. There is no app to install.
  2. Upload a clear photo of your face.
  3. Keep “My outfit photos” and add four outfit photos in the order you want them, or pick a theme pack.
  4. Choose a model and a resolution; the credit price shows before you start. Leave Trend sound on, or pick Silent.
  5. Tick the consent box and press the button. A video takes a few minutes, and you can leave the page while it runs.
  6. Download the 15-second MP4, then turn on the AI-generated label when you post it.

No outfit photos? Use a theme pack

Each pack has four looks written into the template and styled to you. Use one to try the trend; switch back to your own photos when the clothes matter.

PackLooks, in order
EverydayStraight jeans and a white tee; a camel trench over a black turtleneck; a grey hoodie and cargo trousers; a cream cable-knit and navy trousers
AutumnA cream chunky knit and brown cords; a brown leather jacket and dark jeans; a beige trench with a tartan scarf; a black velvet evening look
Night outA black leather jacket and slim jeans; a silver sequinned top; deep red satin with gold jewellery; a white tailored suit
HalloweenA witch; a vampire cape; a skeleton bodysuit; a round pumpkin costume
FormalBlack tie; all white; emerald satin; champagne sequins

Price per video

Credits for one 15-second video, read from the generator. Own outfits and theme packs cost the same, and so do Trend sound and Silent.

Model480p720p1080p
Seedance 2.0 Mini84 ($0.83)180 ($1.78)—
Wan 3.0250 ($2.48)499 ($4.94)997 ($9.87)
Seedance 2.0 Fast180 ($1.78)389 ($3.85)—
Seedance 2.0 Pro304 ($3.01)651 ($6.44)1,634 ($16.18)
Seedance 2.5437 ($4.33)982 ($9.72)1,742 ($17.25)
MiniMax H3—433 ($4.29)—

Dollar figures use the Studio pack ($99 for 10,000 credits, $0.0099 a credit); smaller packs cost a little more per credit. Free credits only pay for Seedance 2.0 Mini; the other models use purchased credits.

Free credits, watermark and refunds

QuestionAnswer
Can I try without an account?No. A free account keeps your photos and videos private to you. No card is asked for at sign-up.
Do welcome credits make a video?Not on their own: the few welcome credits some new accounts get are below the cheapest video (84 credits).
Cheapest way to tryThe one-time $4.99 trial pack: 99 credits, enough for one Seedance 2.0 Mini 480p video.
WatermarkNone, on any video.
Failed videoAll credits back automatically, including when Seedance 2.5’s check rejects a photo.
SubscriptionNone. Credit packs are one-time and last 365 days.

Shop clothes, brand posts and paid ads

  • Product photos in order: upload four product shots as Look 1–4 to show a new drop in the sequence you choose.
  • Show the real product as well: the AI can simplify prints and trims, so pair the video with real photos on the product page and check the details before posting.
  • Label it: TikTok, Instagram and YouTube ask for realistic AI-made video to be marked. Turn the label on and say in the caption that the video is AI-generated.
  • Ads: videos made with purchased credits include commercial use, ads included, with a downloadable certificate; videos paid with free credits do not. Music from a platform’s library is often not cleared for ads, so pick Silent and add audio you are licensed to use.

What this template makes

Clips that follow the same roll onto the rug, overhead fan, four looks and sound as the template this generator uses. We only show clips that match the template, which is why there are two.

Yellow sundress, leather jacket, white crop top, white short set
Four áo dài looks, from everyday to lace

Questions people ask

Put your own outfits under the fan

One selfie, four outfit photos, 15 seconds. No watermark, and a failed video is refunded.