Your full-length photo turned into a tiny figure inside a glass snow globe, on a decorated mantel or at a Christmas market. Face and clothes stay yours. Workflow visible here.
Real workflow: a man in a mustard sweater, light jeans and white sneakers standing on a gravel path in autumn, then the same man as a miniature in a snow globe on a mantel, in the middle of a snowy village, and in a globe held up in red mittens at a Christmas market, red scarf around his neck. He comes out bigger than asked in both globes, and an empty gold plate showed up on the base.
At a glance
The snow globe with you inside comes back every Christmas on TikTok and Instagram. It isn't new, it's a seasonal format that climbs again as December gets closer. The "Christmas TikTok Trends 2026" guide from dromana.app, updated on 21 October, lists it among the AI Christmas formats to post between early and mid-December, and some photo apps already have a snow globe filter. This workflow starts from your full-length photo and makes two images: you in the middle of a small snowy village in a globe on a mantel, then in a globe held up at a Christmas market.
The model keeps the whole person: face, hair, sweater, jeans, shoes. The prompt asks for a figure about a third of the globe's height. In the demo, the man comes out bigger than that in both renders, taller than the village houses on the mantel. At the market, an empty gold plate appeared on the base, and there's hardly any glitter.
To see yourself as a boxed collectible figure, use the figurine photo. To cover your house in lights, see the holiday house decor, and to send your wishes, the greeting card. Both renders cost 14 credits.
These are the files of the workflow above, exactly as the models produced them, unretouched. The starting point is shown when there is one.
Starting point
Render
People who want to do the snow globe trend for their stories or holiday messages, couples and families looking for a different Christmas photo, content creators planning their December posts.
A full-length photo, head to shoes, with a sharp, well-lit face, of you or of people who agreed, and your clothes described in the prompt. The photo is optional.
Use a full-length photo
Head to shoes, standing, sharp face, good light. The figure is whole inside the globe: whatever the photo cuts off, the model makes up.
Describe your clothes in the prompt
Both prompts name the demo's mustard sweater, light jeans and white sneakers. Swap in what you're wearing, or the model mixes the two outfits.
Pick the mantel or the market
The mantel puts you in a miniature village, the globe sitting in a living room. The market shows the globe held out at arm's length, with a red scarf added. Run both, or a single node for 7 credits.
Check the size of the figure
In the demo, the man grew past the third the prompt asked for. If you want him smaller, write "no more than a third of the globe's height, smaller than the houses" and rerun that node.
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
On the mantel, the globe is part of the home decor, with a miniature village around you and one hand reaching for the flakes. At the market, red mittens hold the globe in front of the stalls and you wave at the camera. The first looks like an object sitting at home, the second like a photo taken outside.
Start from a photo where everyone is standing and fully in frame, then write in the prompt how many of you there are and what each person wears. The more people, the smaller each face gets in the globe. Not tested in the demo, and everyone has to agree.
Add the pet to the prompt, sitting at your feet in the snow, with its breed and colour. If it's in the same photo as you, the model has something to copy. If not, it invents one that fits the description. This version wasn't tested here.
Replace the mantel sentence with the spot you want: a windowsill with snow outside, a set table with candles and glasses. The rest of the prompt stays the same. The demo only shows the mantel and the market.
The globe image can be the main picture on a card. For a finished card with text and layout, go through the greeting card and describe the snow globe in the scene. The vertical 3:4 format suits a card well.
Connect a Video node to one of the two renders and describe the motion: a hand shakes the globe, the snow lifts and settles back around you. This isn't shown in the demo, which stops at the two still images, and the video costs extra credits.
The figure is full-length inside the globe. With a cropped photo, the model makes up your legs and shoes, and the jeans or sneakers stop looking like yours.
The prompts describe a mustard sweater, light blue jeans and white sneakers. If you're wearing a black coat, change those lines, or the model dresses you partly like the man in the demo.
Even in miniature, a face and a figure are recognisable. Only use photos of people who said yes, and never a famous person.
No, it's optional. Without one, the model invents the person from the prompt. With yours, it keeps your face, your hair and your clothes, which is the whole point of the trend.
Yes, from a photo where everyone is full-length. Rewrite the prompt with the exact number of people and one line of clothes each. The demo only shows one person, so those versions weren't tested here.
The mantel prompt asked for about a third of the globe's height, and the model made him more than half. Rerun with a stricter size if the scale bothers you.
14 credits, 7 per render on Nano Banana 2.1 in 1K. Rerunning one globe costs 7 credits, or 9 if you switch to 2K.
When you want the globe to fill more of the frame, or a sharper face on the figure. At the size shown in the demo, Nano Banana 2.1 is enough, and Nano Banana Pro costs more per render.
In 1K the image is 896 x 1200 pixels, vertical 3:4: fine for a story or a message, too small for a print. To print, switch the node to 2K, 9 credits per render.
The market render added an empty gold plate, even though the prompt said "no text on the base". You can ask for a name engraved on it, spelled out letter by letter in quotes. That wasn't tested, and a short word is more likely to come out clean.
Both prompts asked for a little glitter, silver or gold. The model went mostly for white snow. For a globe that sparkles, write "lots of gold glitter" and rerun the node.
Put yourself in a snow globe with your photo and AI
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.
Nano Banana 2.1
Flux Kontext
Seedream 5 Pro