The ceiling fan transition is a 15-second video shot from the ceiling: you lie on a rug under a turning fan, and every time a blade sweeps across the lens you are wearing a different outfit. videosi (videosi.io) makes the AI version from one photo of you, with four looks and three blade-covered changes, and you write no prompt.
Key takeaways
- The ceiling fan transition is an outfit-change video: one top-down shot, a fan at the top of the frame, a new look behind each blade.
- It spread on TikTok from the end of September 2026; nobody has confirmed who posted the first one.
- You can film it for real with a phone above a fan, or make it with AI from one full-body photo.
- On videosi a 15-second video on the Standard model at 720p is 13 credits; there is no free generation, and a failed render returns its credits.
- The video is silent, so you add the trend sound or your own audio where you post.
In this guide
- Why does the ceiling fan transition look so clean?
- Are the viral ceiling fan videos real or AI?
- How do people film the ceiling fan trend without AI?
- How does the AI version work from one photo?
- What photo gives the best ceiling fan transition?
- How many outfits should the video have, and which ones work?
- Can I make it with ChatGPT, Sora, Veo or Kling?
- Why does the outfit melt instead of switching?
- Can I use it to show off clothes I sell?
- How do I post it so it does not look fake?
- Where can I keep reading?
Why does the ceiling fan transition look so clean?
Because the blade does the editing. A transition needs a moment where the viewer cannot see the change happen, and a wide fan blade passing close to the lens hides the whole body for a fraction of a second. The camera never moves, the person lies in the same spot and the pose barely changes, so the only thing that differs on either side of the blade is the clothing. The brain reads that as a magic swap rather than a cut.
The setup also helps. A camera pointing straight down shows the outfit head to toe, like a flat lay with a person in it, which is why fashion creators picked it up so fast.
Are the viral ceiling fan videos real or AI?
Both, and it is often hard to tell. Some creators film it for real; many others make it with AI from one photo. A UK student news site covering the trend at the end of September 2026 noted that the AI results looked real enough that some viewers did not realise they were AI (The Tab).
A few tells: in AI versions the room is often the same, with a dusty-blue rug, the edge of a pink sofa and a brushed-nickel fan, because many people use the same template scene. In filmed versions the fan is usually wood or the room is clearly someone's own living room, and the person shifts more between takes.
How do people film the ceiling fan trend without AI?
The filmed method is simple in theory and fiddly in practice. You fix a phone to the ceiling or a tall tripod pointing down past the fan, lie in one marked spot, and record a short take in each outfit with the fan running. In an editor you cut each take at a frame where a blade covers your body, and line the takes up so your head stays in the same place.
What makes it hard: you need a fan with room above it, a safe way to mount the phone, four outfits and a lot of getting up and lying down again in exactly the same spot. A small shift between takes shows as a jump. Some people skip the real fan and lay a stock clip of a turning fan over their footage.
How does the AI version work from one photo?
On the videosi ceiling fan transition page, your photo goes to ByteDance's Seedance video model together with a short top-down reference clip and a written prompt. Seedance 2.0 takes images and video as references (ByteDance Seed), which is what lets the reference clip carry the timing of the blades while your photo carries your face and body.
The prompt has five parts: an identity lock that keeps your face and build the same in every look, the motion reference, the scene and four described outfits, a timeline of eight beats over 15 seconds, and rules such as "the outfit only changes while a blade covers the body". The full prompt is not published, but the page explains each part.
The outfit in your photo is look 1. The other three come from the style you pick: Mixed styles, Glam gowns or Streetwear.
What photo gives the best ceiling fan transition?
A full-body, front-facing photo of only you, in even daylight. The video shows you head to toe for 15 seconds, so a photo cut off at the waist forces the model to invent your legs and shoes, and invented detail is the first thing people notice.
- Keep the face clear: no sunglasses, hats or hair across it.
- Start in something plain. A simple top and jeans make the jump to a gown feel bigger.
- Avoid group photos, mirror selfies with the phone covering the face, and harsh flash.
Use only photos of yourself, or of an adult who has agreed. videosi does not accept photos of celebrities, public figures or minors.
How many outfits should the video have, and which ones work?
Most remakes show four or five looks in about 15 seconds, roughly one blade pass every three to four seconds. Fewer looks feel slow; more than five and the viewer cannot take any of them in.
Contrast is what sells it. Going from casual to formal, or from dark to bright, makes each swap obvious on a small screen. Outfits in the same colour as the rug tend to disappear, which is one reason the blue-rug template pairs so well with reds, pinks and greens.
Can I make it with ChatGPT, Sora, Veo or Kling?
You can try with any video model that accepts an image, but a text prompt alone rarely gets the timing right. The hard parts are keeping the camera locked overhead, keeping the blades rigid and turning one way, and making the outfit change only while a blade covers the body. Without a motion reference the model usually morphs the clothes in plain view or drifts the camera.
Expect several tries, each costing credits and minutes. The ceiling fan transition AI generator on videosi has the reference clip, the scene and the rules built in, so you upload one photo and pick a style.
Why does the outfit melt instead of switching?
Video models like smooth change. Asked to go from a dress to a jacket, they will happily blend one into the other over a second, which looks like melting fabric. The fix is to tie the change to a moment the viewer cannot see: the prompt says the outfit may change only while a blade fully covers the body, and the reference clip shows exactly when that happens.
If a result still melts, the usual cause is a photo where the body is small in the frame or partly hidden, so the model is unsure where the clothes are. A failed, timed-out or blocked render on videosi returns its credits automatically.
Can I use it to show off clothes I sell?
Not item by item on this page. The three styles describe generic outfits, and logos and text on clothing are left out on purpose, so the video cannot reproduce a specific product you upload. Fashion sellers who want their own pieces in each look usually film the trend for real, as described above.
How do I post it so it does not look fake?
Add a sound inside TikTok, Reels or Shorts; there is no single official song, and many remakes reuse one popular original sound. Turn on the platform's AI-generated label, which major platforms ask for on realistic AI video, and do not claim it was filmed. Platform rules change, so check the current guidelines.
Pick 9:16 for TikTok, Reels and Shorts. 1:1 suits a square post, and 16:9 a wide cut.
Where can I keep reading?
- All videosi effects, from hugs to transformations.
- Pricing and credit packs: packs start at $6.99 for 10 credits and never expire.
- Which photo works best for AI video effects.
Ready to try it? Open the ceiling fan transition page, upload one full-body photo, pick a style and download your 15-second outfit swap.