The UNO reverse car transition is a 12-second clip in which a person lifts two cards until they fill the screen, lets them fall, and finds a car parked behind them that was not there a moment earlier. It works because the cards hide the cut. It is hard to copy from a tutorial because three details go wrong again and again: the digits on the cards, the moment the car appears, and the face. videosi builds all of it in, so you upload one photo on the UNO reverse car transition page and download the finished clip.
This guide explains the trend rather than teaching you to rebuild it by hand. You will see what happens second by second, why the cards are the whole trick, why the recipes that float around usually disappoint, and what you get when the hard parts are already done for you. Tutorials for the clip began circulating on YouTube and Instagram in October 2026, usually with the song "Unforgettable" underneath, so most people meet the trend first as a finished video and only later as a how-to.
In this guide
- What the UNO reverse car transition is
- What happens in the 12 seconds
- Why the cards hide the cut
- How these clips are made
- Why copying a prompt from a tutorial usually fails
- What videosi does instead
- The photo, the car and the cost
- Make yours
- Continue reading
- Questions
What is the UNO reverse car transition?
The UNO reverse car transition trend is a 12-second video in which a person holds up two playing-style cards, covers the camera with them, and reveals a car behind them when the cards drop. It is built around one trick. Someone stands close to the camera in an empty car park, lifts the cards until they fill the screen, and lets them fall. Behind them is a car that the first three seconds made a point of not showing.
Then comes the victory lap: a thumb over the shoulder toward the car, a hand beside the mouth like a shared secret, and the cards once more to close the clip.
The two cards are a 2 and a 4, which is why some people search for the trend as the "24 UNO car" clip. Nobody has settled on what the number means. Read together the cards say 24 and that is as far as any agreed meaning goes. What the cards do in the video is practical, and that is the interesting part.
The format spread because it asks so little of the person in it. There is no car to borrow, no location to find and no editing skill to learn. The clips are generated by an AI video model from a single photo of the person, a short reference clip that supplies the movement, and a written prompt that describes the car and the place.
What happens in the 12 seconds?
The clip has six beats, and each one has a job. Here they are in order.
| Time | Beat | What you see |
|---|---|---|
| 0 to 3 s | The empty lot | You are close to the lens, looking away, then you turn to camera. Nothing is parked behind you. |
| 3 to 4.5 s | Cards toward the lens | One hand brings up the 2 and the 4, fanned out, and pushes them forward until they take over most of the frame. |
| 4.5 to 5 s | The hidden cut | Your arm crosses the lens. For a handful of frames there is only sleeve and card. |
| 5 to 7.5 s | The reveal | The cards fall away and the car is simply there. You look back at it and point with your thumb. |
| 7.5 to 10.5 s | The flex | Back to camera, one hand beside your mouth, a single nod. |
| 10.5 to 12 s | Cards out | The cards snap up in front of the lens a second time and the clip ends, ready to loop. |
The opening is the setup: the shot lingers on an empty space long enough for the viewer to notice it. The cards are the trick. The payoff is the last seven seconds, where you react to the car with a glance, a thumb, a shared secret and a nod, so the reveal reads as a small story with a punchline.
Why do the cards hide the cut?
The cards hide the cut because for about half a second they cover the lens, and that half second is where an empty lot becomes a lot with a car in it. When the cards drop, the car is already in place and looks as if it had been there all along.
Hiding a change behind something that fills the lens is an old idea in video. What makes it work here is the way the clip is structured around it. The opening tells the viewer the space is empty, the arm sweep gives the cut a reason to exist, and once the cards fall the person acts as if the car is old news.
The reveal only lands if the car is fully present the moment the cards drop. A car that is half there or still fading in is a visible error.
How are these clips actually made?
These clips are generated, not filmed, and they rest on three ingredients. Each one is also a place where a copy can go wrong.
- A single photo of the person. It supplies the face, the hair and the outfit. Whatever you wear in the photo is what you wear in the video.
- A short reference clip. It supplies the movement: the turn, the card push, the thumb, the nod, and where the camera sits.
- A written prompt. It describes the car, the place, the time of day and the order of events.
A tutorial that shows only the third ingredient is showing one-third of the recipe. The other two are the photo, which is different for everyone, and the reference clip that supplies the movement. This is the root of most failed attempts, and it leads to the next question.
Why does copying a prompt from a tutorial usually fail?
Copying a prompt from a tutorial usually fails because the text is only part of what produced the clip you admired, and because a video model does not follow written instructions frame by frame. Three failures show up more than any others.
The card digits come out mirrored
The cards are the one place where precision matters most. They are small flat objects, they move toward the lens, they tilt and they carry a number that must read correctly: a 2 and a 4. A video model paints each frame as an image. It has no rule that says a 2 must stay a 2 as the card turns, so digits slip into mirror images, change shape or become something that only resembles a number. A card that shows the wrong digits breaks the joke, because the 2 and the 4 are what make it recognisable as this trend.
The car appears too early
The whole clip depends on the car not being visible until the cards cover the lens. A written timeline is a request, not a guarantee. A model may let the car drift into the background in the first seconds, or fade it in while the cards are still moving. Either way the viewer sees the car arrive, the hidden cut is no longer hidden, and the reveal is gone.
The face changes
You turn to the camera, your arm crosses the lens, and your face disappears for a few frames and comes back. Every time a face leaves the frame and returns, the model has to redraw it. Details drift: the hairline moves, the nose shifts, the expression changes. A reference clip adds another risk, because it shows a different person. Unless the instructions are very clear about what to take from it (the movement) and what to ignore (the face, the clothes, the card artwork), a model can borrow from it.
Every attempt has a cost
A 12-second render is not a quick test, and getting it right by trial and error means paying for several failed clips. That is a poor trade when you want one specific, well-known clip.
What does videosi do instead?
videosi (videosi.io) makes the UNO reverse car transition from one photo, with the movement, the card design and the instructions already built in. You do not write a prompt, find a reference clip or design a card. The three things you choose are the photo, the car and the shape of the video.
- The movement is built in. The turn, the card push, the thumb over the shoulder and the nod follow a fixed reference, so the timing is the same each time.
- The cards are ours. They use a generic design, not UNO's, and the 2 and the 4 are specified as digits that must stay readable and unmirrored. UNO is a trademark of Mattel, and videosi has no connection to Mattel.
- The rules of the reveal are built in. The car must not show before the cards cover the lens, and it must be fully there when they drop.
- Your face is protected. The settings put your likeness first and tell the model to take only movement from the reference, not another person's face, clothes or card art.
The result is a 12-second MP4 at 480p in 9:16 or 16:9, with no watermark and no sound. It is silent on purpose: add the trending sound inside TikTok, Instagram or YouTube Shorts, where the platform holds the music licence.
Honest limits apply. The page renders three built-in cars and not your own car. Resolution is 480p. And an AI render can still come out imperfect, since likeness is never guaranteed. If a render fails, the credits return automatically. If a completed result is clearly broken, such as an empty or corrupted file, write to us and we will return them.
Which photo, which car and what does it cost?
Pick the photo
The clip is mostly upper body and one raised arm, so a photo with your face to the camera and your shoulders in frame works best. If your hands are visible, so much the better. A photo cropped tight to the face leaves the model to invent the rest, and hands are where invented detail shows first. Leave hats, helmets and sunglasses off. Anything that covers your face makes the likeness drift during the turn to camera.
Dress for the car you pick: contrast makes the reveal readable on a phone, so avoid a red top with the red supercar. The night scenes are the most forgiving if your photo was taken indoors.
Pick the car
| Car | Place and time |
|---|---|
| Red supercar | Open outdoor parking area at sunset |
| Muscle car | Empty concrete lot at night under floodlights |
| JDM tuner | Rooftop car park at night, wet asphalt and neon |
Know the cost
The UNO reverse car transition costs 100 credits per video, ten times a standard effect, because the clip is 12 seconds long. There are no free generations.
| Pack | Price | Credits | UNO videos |
|---|---|---|---|
| Starter (one per account) | $2.99 | 30 | 0 (three standard effects) |
| Small | $9.99 | 120 | 1, with 20 credits left |
| Large | $24.99 | 350 | 3, with 50 credits left |
Full details are on the pricing page. You must be 18 or older, and photos of celebrities and public figures are not accepted. Your photo and the result are deleted within 30 days.
Make your own UNO reverse car transition
If you want the clip and not the experiment, go to the UNO reverse car transition page on videosi, upload one photo of yourself, choose a car and a format, and press Generate. You can close the page while it renders; the finished video is saved in My creations.
Continue reading
- More viral effects, including the Earth zoom out video.
- All videosi effects, from hugs to anime.
- Pricing and credit packs.
Questions
Why does the car show up before the cards in some UNO reverse car videos?
Because a written timeline is a request the model may not follow to the frame, so the car can drift in during the opening seconds. The reveal only works when the car is absent until the cards cover the lens. videosi builds that rule into the 12-second UNO reverse car transition, so you choose a photo and a car and do not manage the timing.
Why are the numbers on the cards backwards in my AI UNO video?
A video model paints each frame as an image and has no rule that a 2 must stay a 2 while a small card tilts toward the lens, so digits mirror or warp. videosi uses its own card design with the 2 and the 4 specified as readable, unmirrored digits, so you do not have to correct the cards yourself.
Why does my face look different in my UNO reverse car video?
Your face leaves the frame behind the cards and has to be redrawn, and a reference clip showing another person can leak into the result. videosi puts your likeness first and takes only movement from its reference. Likeness is still not guaranteed, so choose a sharp, front-facing photo with your shoulders in frame.
How many UNO reverse car videos can I make with the Small pack?
One. The Small pack is $9.99 for 120 credits, and each UNO reverse car transition on videosi costs 100 credits, which leaves 20 credits, enough for two standard 10-credit effects. The Large pack is $24.99 for 350 credits and makes three UNO videos. The $2.99 Starter pack has 30 credits, so it covers standard effects only.
Can I make the UNO reverse car trend with a group photo?
The effect is built around one person, so upload a photo of a single adult with your face and shoulders in frame. A crowded photo gives the model more faces to keep consistent through the arm sweep and the turn. videosi accepts JPG, PNG, WEBP and HEIC files up to 10 MB, and deletes your photo within 30 days.