Mini Me AI Video — A Tiny You Meets Ice Cream
A tiny you on a sunny window ledge: a cheek tap, a grumpy face and a surprised look at ice cream. Watch our original one-character example and create a ten-second video with sound.
The idea behind the effect
What is Mini Me AI?
Mini Me AI turns someone's appearance into a miniature character in a short video. A tiny person sits on an everyday surface and interacts with a normal-size hand and an ice cream cone. The window frame, wood grain and fingers establish scale, while the character's changing expressions give the little figure a personality.
Mini Me can also describe miniature portraits and collectible figures; it does not refer to one universal sequence. This version follows a cheek tap, an annoyed reaction, the discovery of ice cream and a happy ending. One miniature adult appears throughout. The offscreen helper contributes hands rather than a second full character.
What does our one-character example show?
A fictional adult woman with wavy brown hair, a gray top and dark trousers sits on a sunlit wooden window ledge. After a finger touches her cheek, she folds her arms. When a cone approaches, her eyes widen and her mouth opens. She takes the cone and leans toward the ice cream, receives a light head pat, then stands and walks while holding it.
Our saved example is approximately ten seconds, 720p and vertical 9:16, with generated audio. The character has fairly natural miniature proportions, and cream on the face is not a prominent detail. We generated a new scene using an original fictional project portrait, without reusing another creator's footage or soundtrack. Your appearance and movements can differ from the preview.
Why do the reaction and handoff matter?
The crossed arms establish annoyance before the food arrives. Looking at the cone, opening the eyes and then reaching for it makes the change of mood readable. Seeing the cone before taking it, and holding it before eating, gives the clip a small story rather than a collection of unrelated gestures. One character and a consistent outfit help viewers follow that story.
The contrast between a large hand, a separate cone and a tiny person creates the visual joke. The scene instructions request a stable handoff followed by eating and a few final steps. AI can still change the grip, character scale, finger contact and timing; every beat is not guaranteed to match the preview.
Which photo should I upload?
Use one sharp adult portrait with the face, hairstyle and upper body visible. Even lighting and a front-facing or gentle three-quarter angle make the person easier to recognize. Avoid groups, covered faces, heavy filters and a very distant subject. Use a photo you have permission to process.
The image supplies appearance, while the window ledge, cone and movement are generated. You do not need a window photo, an ice cream prop or a recording of yourself acting. This version supports one adult human subject; pets, groups and alternative interaction stories are not available in this template.
Sound and likeness
The template requests fresh playful instrumental music, small interaction sounds and a wordless laugh. It does not offer a named song, spoken dialogue or a lip-sync performance. Audio details, action timing and facial likeness can vary. Hands, food contact, cream placement and consistent scale also need checking in a finished result.
Create your Mini Me video
Sign in with a verified email account and choose Mini Me in the creator. Upload one portrait, check the ten-second, 720p, 9:16 specification and 100-credit cost, confirm photo permissions and submit once. Follow the task in Your videos and download the private result when ready. Watch the preview and check current availability before purchasing credits.
Related scenes and references
Try Window Dance for another change of expression, Street Light for a physical comedy scene, or Tiny Desk AI for a character performance.
For related interpretations, see MakeMeme's Mini Me introduction and MakeViral's Mini Me page. Vidu's subject-reference documentation describes appearance references and audio controls. Examples and product settings differ between services; these references do not establish identical models or guaranteed outcomes on this site.
From input to output
How to use it
1. Watch the one-character preview and check that Mini Me is available to create. Sign in with a verified email account.
2. Upload one clear adult portrait in PNG, JPEG or WebP format, up to 10 MiB. Your photo does not need to match the window ledge background.
3. Check the 10-second, 720p, 9:16 specification and 100-credit cost. Confirm photo permissions, then submit once.
4. Follow the task in Your videos and download the private result when ready. Check an uncertain task or contact support before submitting again.
Good to know
Best for
Playful portrait stories, miniature-character clips, ice cream reactions and creative videos for friends. Present the result as an AI-created scene.
Limits & considerations
One adult portrait only. AI regenerates the setting and motion; likeness, proportions, hands, cone handoff, cream details and walking can vary. Generated audio does not promise a named song, dialogue or accurate lip sync. The character starts miniature; no gradual shrinking sequence is included. One complete video costs 100 credits. Check the preview and current availability before buying.
Questions about this effect
What is Mini Me AI?
It turns a portrait into a miniature adult on a window ledge, using a cheek tap, changing expressions and an ice cream cone to tell a tiny-character story.
How many characters are in this template?
One miniature adult appears throughout. Offscreen hands provide interaction without a second full character. You only upload one portrait.
Do I need a video of myself acting?
No. One clear portrait supplies the appearance reference. The window ledge, cone, expressions and motion are generated.
Does the person gradually shrink?
This version begins with an already miniature person. The focus is scale contrast and expressive interaction rather than a shrinking transition.
Will it look exactly like my photo, and can I use pets or groups?
The model references appearance but cannot guarantee an exact match. This template is designed for one adult human portrait; pets and group photos are not supported.
Does it have sound, and can I choose a song?
Yes, it includes generated audio and requests instrumental music with interaction sounds. Song, dialogue and voice selection are not available, and audio and action timing can vary.
How long is it and how many credits does it cost?
One private video is approximately ten seconds, 720p and vertical 9:16, and costs 100 credits. Check current creation availability before buying.
Keep exploring.
All effects →Window Dance AI Trend — Dance Behind the Window
A warm window, a private dance and an awkward discovery. Watch our original Window Dance example and create a twelve-second nighttime scene from one portrait.
View effect guideStreet Light Trend AI — Streetlight Slip Video
A person slips away beside a street lamp, then returns close to the camera with a calm smile. Watch our original five-second example and learn how a portrait becomes a playful night scene.
View effect guideTiny Desk AI Video: The Desk Concert Trend
Explore the Tiny Desk AI desk-concert trend and replace one to three characters with photos of people or pets. Each complete private video is about 18 seconds, with the template soundtrack, for 100 credits.
View effect guide