Video trend
The clip already tells a full story in 17 seconds: a friendly walk, a helpful button press, a betrayal and a parting gift. Your caption and casting decide who the joke is on. This page covers who to put in each role, captions that work with the betrayal, the cover frame and how to label the video.
The classic version is your dog as the Prankster and you as the Victim: the dog you trust with everything kicks you into a lift. Swap them and you become the villain kicking your dog in. Two friends work as a revenge clip, and two pets as a sibling rivalry. Whoever you cast, use only photos of people who agreed to appear.
The reference clip's caption, "bro trusted the golden retriever for ONE lift ride", shows the shape that works: set up trust, let the video deliver the twist. Other directions: "He pressed the button for me. I should have known.", "Taking my shoes off was the mistake", or naming your dog: "Biscuit has never forgiven me for the vet".
The opening frame, with the Prankster's paw or hand on the Victim's shoulder as they walk to the lift, makes the strongest cover because it looks innocent. The kick at about 11.4 seconds is mostly motion blur and makes a poor still.
The ceiling and the upper walnut panels are empty for most of the clip, so on-screen text near the top stays clear of both characters. Avoid the lower third: the sneakers sit on the floor there from about 9 seconds and the Victim ends the clip on the lift floor.
Your video keeps the reference's original sound: short spoken lines and the sound effects of the smack, the kick and the shoes landing. There is no song under it, so the hits land on their own; if you add music in an editor, keep it low enough that the kick still lands.
Say in the caption or with the platform's AI label that the video is AI-made. If you post the reference next to your version, say which one is yours.
Add a photo for each role shown below. All roles need a photo.
Upload the prankster and the victim, one subject per photo: a pet, yourself, a character or someone who agreed to appear. Full-body photos with the face, legs and paws or hands visible work best.
Choose 480p, 720p or 1080p and review the quote. It includes both photo preparations and the full video.
Watch that paws stay paws when the prankster presses the button and throws the sneakers, that the two never swap at the kick, and that the close pass at the end keeps the prankster's face, then download your portrait MP4.
The reference prankster is a golden retriever walking upright, but your photo decides: your own dog, a cat, a friend or a character. An animal keeps its species, face, coat and tail; a person stays a person.
The victim gets smacked, pulls off a pair of white sneakers at the lift, is kicked in and ends up sitting on the lift floor. Put yourself, a friend or a second pet in that role. The white sneakers are the one prop the template adds.
The marble lobby, the walnut panels and the gold lift stay as they are. The template follows the reference beat for beat in one continuous 17-second handheld take, with the original spoken lines and hit sounds.
The preview is the AI reference clip the trend comes from, with its watermark removed. It shows the motion and timing your video follows, not an Oakgen result.
It is a 17-second portrait AI clip: a golden retriever walks upright beside a man to a gold lift in a marble lobby, smacks him on the neck, presses the call button for him, waits while he bends over to pull off his white sneakers, then kicks him into the lift and throws the sneakers in after him before strolling past the camera. Oakgen's template puts your two photos in those two roles and keeps the lobby, the timing and the sound.
We did not find an established name for this clip, so Oakgen names it by what happens. The posts around it use words like lift prank, elevator prank and dog prank; lift and elevator mean the same thing here.
Yes. Both roles are replaced in every video, so the prankster and the victim each need a photo before you can generate. Neither performer from the reference is kept.
Yes. The prankster keeps the uploaded animal's species, face, coat and markings and walks upright on its hind legs, using its front paws for the smack, the button and the sneakers. Bodies very different from a tall retriever vary more, so review the video before sharing.
Yes. Either role takes a pet, a person or a character photo and keeps its species. A person prankster does the walk, the smack and the kick with hands and feet. A pet victim still takes off the white sneakers at the lift, which is the oddest moment to check.
Their own outfit from the photo, plus one template prop: plain white sneakers that come off at the lift and get thrown in after them.
No song. The video keeps the reference's original sound: short spoken lines and the hit sounds of the smack, the kick and the sneakers landing. Your photos do not change the voices.
Your own pets, yourself, or people who agreed to be in the video. Do not upload photos of celebrities, public figures or anyone who has not consented, keep it obviously a joke and label AI videos clearly when you share them.
About 17 seconds in portrait 9:16, one continuous take from the walk to the lift to the prankster walking out of frame.
No. It is the AI reference clip the trend comes from, with its watermark removed, not an Oakgen result. Review paws, hands and faces in your video before sharing.
3,596 credits per 720p video. The quote includes both photo preparations and the complete video. Check it before generating.
Plan a story, keep your character recognizable, and make the finished clip your own.
Five ways to cast the Dog Elevator Prank: your dog and you, you and a friend, two pets, characters and the role swap, with what each changes and what to check.
Make the Dog Elevator Prank AI video step by step: pick the Prankster and Victim photos, read the quote, check the kick and the shoe throw, and post it.
Upload 2 photos. Everything else is set.