Video trend
The reference customer is a golden retriever, but the Window customer slot also takes a photo of a person. The scene, sound and order of gestures stay the same; you become the one at the glass, supervising a slice of tiramisu with total gravity. Here is what changes when a person takes the role.
The template keeps the uploaded character’s species, so a person stays a person and makes the gestures with hands instead of paws: a thumbs-up at the server, both hands held out to size the slice, a hand to the chest, a hand on the head at the box and a raised hand before the final stare. You lean on the same window ledge outside the same shop. The seller’s hands, the tray, the lettering on the glass and the street do not change.
With a dog, the humour is an animal acting like a picky customer. With a person, it becomes an adult treating one slice of cake like a major purchase: guiding the cut to the millimetre and visibly despairing at the box. It plays as self-parody, so it suits creators who already post about food, diets or being particular, and it suits a friend you want to tease, with their permission. These are editorial suggestions, not tested results.
Use one clear photo of the person facing the camera, face fully visible, in even light. Head and shoulders with some of the chest shows enough of the outfit, and your clothes in the photo are what you will wear at the window. Avoid sunglasses, hats that shade the eyes and group photos; crop down to the one person. A neutral or serious expression in the photo fits the deadpan role better than a big smile.
Only upload a photo of another person if they have agreed to appear in the video. The result shows them at a shop they never visited, making gestures they never made, so label it as an AI video when you post it. Do not use photos of public figures or anyone who has not agreed.
Captions land best when they name the stakes. Original examples to adapt: “Me ordering one slice like it’s a mortgage” over the cut, “No, a bit to the left” as the hands size the slice, and “They put it in a box like THAT” on the facepalm. Keep text above the tray and away from the face, and leave the final stare clean. These lines are suggestions for your editor; the template does not add text.
A two-part post works well: the original reel or your pet’s version first, then your own version cut in at the same beat, such as the thumbs-up at about 10.25 seconds. Because both versions follow the same timing, the gestures line up across the cut. Keep the original sound under both and label each part honestly, for example “Original reel” or “My dog” and “Me”.
Choose the roles you want to change. Add a photo for each selected role.
Upload one clear photo of your pet or yourself, facing the camera with the face, ears and coat or hair visible. Upright or sitting pets work best.
Choose 480p, 720p or 1080p and review the credit quote. It includes the video plus the photo and scene preparation for your customer.
Watch the paw or hand gestures, the facepalm and the final stare, then download your portrait MP4 with the original sound.
One photo casts the customer outside the shop glass. A dog stays a dog and a cat stays a cat, with its own coat and markings; a person works too.
The full 18-second take keeps the thumbs-up while the slice is cut, the measuring paws, the facepalm at the box and the final stare.
The camera stays behind the counter with the tiramisu tray, the seller's hands, the gold window lettering and the street behind.
The preview is @m.a.n.g.o__o's AI reel captioned "Dead Serious Tiramisu". It is a source reference, not an Oakgen-generated result. The original dog's name on its bib is not copied into your video.
It is a fixed shop-window scene: a customer stands outside the glass and, with a completely serious face, directs where the tiramisu gets cut, measures the slice, facepalms at the box and stares. Tiramisu Dog is Oakgen's descriptive name for the supplied AI reel.
Yes. The reference is @m.a.n.g.o__o's AI reel captioned "Dead Serious Tiramisu", with a golden retriever as the customer. In Oakgen, your uploaded photo replaces that dog.
Yes. The template keeps the uploaded animal's species, coat and markings and adapts the gestures to its paws. Results vary with unusual body shapes, so review the video before sharing.
Yes. A person stays a person and makes the same gestures with their hands at the window. Upload a clear photo with the outfit you want.
No. The name on the original dog's bib is not copied. An animal without clothing may get a plain white bib with no text.
One. The single photo casts the customer for the whole clip. The seller's hands, tray and street are part of the fixed scene.
The output is about 18 seconds in portrait 9:16 format, matching the full reference.
Yes. The final MP4 keeps the reference's street and serving sounds. There is no speech or song to replace.
No. It is the attributed source AI reel, which is low resolution. Generated likeness and paw contact can vary; review your result before sharing.
3,869 credits per 720p video. The quote includes one character preparation, three scene references and the full video. Review the selected resolution's quote before generating.
Plan a story, keep your character recognizable, and make the finished clip your own.
Turn an 18-second Tiramisu Dog result into a post: captions timed to each paw gesture, sound choices, shorter cuts and honest labels for the source reel and yours.
Choose the customer for Tiramisu Dog. Match a pet or person to the deadpan joke, pick one photo, and know what to check when the slice gets boxed.
Upload 1 photo. Everything else is set.