Christmas AI Photo Prompts

What a Christmas photo prompt needs

Type «make this a Christmas photo» under a selfie and the model reaches for every holiday sign it knows: a tree, snow, a fireplace, a red hat, all lit the same flat way. The picture reads as a collage, and since nothing protects your face, it may come back slightly redrawn. A request that ends up on a card settles five things first.

1. Which photo, and who's in it. «Use the attached photo of the two of us» marks this as an edit of your file and counts the people, so nobody is added or merged into a neighbor.
2. The hold list. Every face, skin tone, hair and body, «exactly as they are», plus glasses if you wear them. For a pet, the markings and fur color: a white patch on the chest is what makes it your animal.
3. One setting, with its light. By the tree, outside a snowy cabin, a 1950s front room. Pick one and say where the light comes from, since holiday light is mixed: warm lamps indoors, cold blue snow light outside. «Warm lamp from the left, tree lights as soft bokeh behind» is something to build on. «Festive lighting» isn't.
4. The clothes, in one line. Matching sweaters, a wool coat and scarf, a deliberately awful knit. Clothes are the easiest thing to change, and they do most of the work of making a picture read as December.
5. Where it's going. 4:5 suits a card and a phone feed. If you'll add a greeting, say which part of the frame stays empty. Then close with no text, no watermark, no extra people.

Put together, a request you can type from memory:

Use the attached photo of [who is in it]. Keep every face,
skin tone, hair and body exactly as they are. Change only the
clothes, setting and light: [clothes], [one setting], [light
source and direction]. Aspect ratio [4:5].
No text, no watermark, no extra people.

Seven prompts on this page are that request filled in for one look; the other five build a scene from nothing. For restyles outside the season, the Gemini photo prompts use the same method.

Which Christmas prompt for which picture

Pick by what the picture is for, then check the last column: a snow globe needs your whole body in the shot, a card background needs nothing.

The product prompt invents an item from your description, fine for a mock-up and wrong for a real listing. For a photo of something you actually sell, start from Product Photo From Snapshot and add the pine and ribbon.

A card request, before and after

We wrote this request, and the version under it, to show what the five points change.

make this a christmas card, cozy and festive

«Cozy» and «festive» describe a feeling, so the model supplies every detail itself. Nothing says the faces stay. There's no frame, no light and no empty space, so a greeting would land across someone's sweater. And with no cleanup line, it may write «Merry Christmas» in for you, spelled however it likes.

Use the attached photo of the two of us and our dog. Keep both
faces, skin tones, hair and bodies exactly as they are, and keep
the dog's markings and fur color as photographed. Change only the
clothes, setting and light: matching forest-green knitted sweaters,
sitting close together on a rug in front of a decorated tree, the
dog lying across both our laps. Warm lamp light from the left, tree
lights behind falling into soft round bokeh, gentle fill so neither
face sits in shadow. Leave the top third of the frame empty.
Aspect ratio 4:5. Photorealistic.
No text, no watermark, no extra people, no extra animals.

The count now includes the dog, and so does the last line. The light is one lamp plus the tree, which is what a living room has at night. «Neither face sits in shadow» matters because a tree-lit room is dark, and darkness is where a face loses detail first. The empty top third is for your greeting, added in a card app. This is the Family Holiday Card Edit adapted for two people and a pet; the couple photo prompts have more settings for two.

Follow-up lines that fix a holiday edit

When the first result is close, reply in the same chat. The model still has your photo and request, so one line usually beats a rewritten prompt.

Where Christmas edits go wrong

Holiday pictures slip in a few recurring places. Check them at full size before anything goes to print.

• Every extra face is another chance to drift. A solo edit has one face to hold; a group card has four or five. Check each one zoomed in, and fix a single person with the follow-up above.
• Mixed light shifts skin. Firelight, tree lights and blue window light each tint a face, and all three together can tip skin orange on one side and gray on the other. That's why each prompt here names two sources at most.
• Hands and props merge. A mug in gloved hands, a gift on a lap: wherever two things touch, they can blend, or a finger appears. Give hands one simple job and check them first.
• Decorations grow fake writing. Stockings with names, labels on ornaments, a sign over a stall. Winter Market Street asks for no readable signs; add that line wherever lettering creeps in.
• The snow globe shrinks your face. Snow Globe Portrait turns you into a figure a few inches tall, and a face that small holds very little detail. It's a fun post. For the card itself, use a portrait-sized prompt.

One rule of ours: these prompts are for adults, couples, friends and pets, in photos of people who've agreed to be edited.

Gemini and ChatGPT for holiday photos

Every prompt here runs as typed in both. These details come from Google's and OpenAI's own help pages and are current as of October 2026; both change often.

• Gemini on a free account. Google lists image generation with Nano Banana 2 for every Gemini plan, including no plan at all. Limits are compute-based and refresh every five hours up to a weekly cap, and Google says media generation uses more of that budget. Start from a finished prompt so the allowance goes on the card.
• Print size in Gemini. Google lists downloads at 2K with a Google AI plan and 1K without one. Check that against what your print shop asks for.
• Several photos at once in Gemini. Nano Banana 2 accepts multiple reference images. If the Gemini model is set to Flash-Lite, it uses Nano Banana 2 Lite, which Google says isn't designed for that, so switch to Flash or Pro before you combine photos of different people. Editing an uploaded photo requires you to be 18 or older.
• ChatGPT. OpenAI lists ChatGPT Images on all tiers, free included. Upload a photo and describe the change; the aspect-ratio picker sets the shape. For a one-spot fix, like a hat that came out wrong, the selection tool marks just that area. OpenAI notes that highlights aren't always precise and edits can spread past them, so check the faces again afterwards.

Plan prices are in our Gemini and ChatGPT reviews. For pictures built from scratch, see the AI photo prompts; for October, the Halloween photo prompts.