Your selfie as a Christmas elf: in Santa's workshop, or in the full costume by the tree. One photo, 38 credits, flaws shown.
Real workflow: a man in a navy sweater against a grey wall, then the same man as a green elf tying a ribbon in the workshop, and in the full costume with striped tights in front of a lit Christmas tree. A wedding ring appeared in both pictures.
At a glance
Every December, people turn themselves into Christmas elves from a single photo. The viralvideoai.app tracker lists the AI elf dance among its current trends (page dated 2 November 2026) and counts 23 million views for it, a number we repeat without having been able to check it. The most watched version is a video, but it all starts from a picture, and that picture is what this workflow makes.
It gives you two images. The first puts you in a green tunic in Santa's workshop, tying the ribbon on a present. The second shows you head to toe, in striped tights and belled slippers, by the living room tree. Nano Banana Pro keeps the demo face well, but it added a wedding ring in both pictures, and the dance had to be redone: the first try had three legs.
For a calmer holiday picture, see the Christmas photo from your selfie, and for a round picture to use as your avatar, the Christmas profile picture. If you like dressing up, the fairy transformation follows the same recipe. Both renders cost 38 credits, see pricing.
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 recreate the Christmas elf trend from a selfie for TikTok, Instagram or a greeting card, parents making a funny picture for the family, coworkers livening up the office Christmas party.
A sharp, front-facing, well lit selfie with a bare head, of you or of someone who agrees, and a description of their face in the prompt (beard, glasses, haircut). The photo is optional.
Take a well lit selfie, facing the camera
The hat covers the top of your forehead and your ears change shape. That leaves your eyes, nose and mouth to recognise you by, so they need to be sharp.
Describe your face in the prompts
The prompts talk about a clean-shaven man with short hair. Swap in your beard, your glasses or your long hair, or the model may remove them.
Run the workshop first
The waist-up portrait holds the face better than the full-length shot. If it does not look like you, change your selfie before paying for the second image.
Check the hands and legs
On the demo, a ring appeared and the first dance had three legs. Look at the picture full size before you post it, and rerun if a limb is extra.
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
Plug several photos into the node and describe each person in the prompt, with matching costumes for everyone. The more faces, the more likely the model is to make two of them look alike. Not tested on the demo, which starts from one photo.
Replace "the man in this photo" with "the woman" or "the child" and adjust the face description. The costume stays the same. For a child, keep the picture for the family and ask the parents before posting it anywhere.
The prompt sets a green tunic with a red collar. Write a white and silver costume, or striped elf pyjamas, if you want to step away from the classic. Keep the pointed hat and ears: without them, nobody sees an elf.
The workshop picture leaves room above the hat. Add a line of text in quotes, "Merry Christmas" for example, or drop the picture into the greeting card template. Nano Banana Pro writes less cleanly than a text model.
This is the most watched form of the trend. Start from the tree picture with a Video node and describe a simple step, hips swaying and arms going up. Complicated moves often give limbs that blend together. Not tested on the demo.
In the stories, an elf fits in your hand. Ask for "tiny, sitting on a branch of the tree next to a bauble" for a funnier picture. The face becomes very small and harder to recognise, so pick a very sharp selfie.
The model has to replace whatever covers your head with an elf hat, and it then invents your hair and forehead. A bare head leaves it less to guess.
The demo's first try, one knee raised and arms out, came out with three legs. A standing pose with both feet on the floor works much better.
Without instructions, it added a wedding ring in both pictures. If someone else gets the photo, that detail can raise questions: say what you really wear.
Yes, by plugging a Video node into the tree picture and describing the dance step. The demo stops at the two images, and the video costs extra credits.
The model allows it, but we only tested an invented adult. Do not post a child's photo without their parents' agreement, and keep the picture within the family.
The model added a wedding ring in both demo pictures without being asked. Write "no ring, no jewellery" in the prompt if you do not wear any.
38 credits, 19 per render on Nano Banana Pro. Rerunning one image costs 19 credits, like the dance redone on the demo.
Yes, by switching the node to Nano Banana 2.1 at 1K, 7 credits per render. The face will be less faithful, especially in the full-length shot. The demo runs on Nano Banana Pro.
1792 x 2400 pixels, portrait, enough for a story, a phone wallpaper or a sharp 4 x 6 inch print.
We do not know who started it this year. viralvideoai.app lists it among its current trends in early November 2026, and insMind writes that it comes back every December. Imaginode has no link with these sites.
The photo is optional in the canvas, but without it the model invents a man. For an elf with your face, you need your selfie.
Turn yourself into a Christmas elf with your photo and AI
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.
Nano Banana Pro
Flux Kontext
Seedream 5 Pro