Christmas AI Photo Prompts: Family, Pets & Product Shots

Christmas AI photo prompts are everywhere by late October: matching pajamas by the fireplace, the dog in a Santa hat, the "turn my family photo into a card" trend. Most lists stop at the prompt. Very few show what it produced, and almost none cover the third thing people actually need in December: product shots for a holiday shop.

So we ran all three. Sixteen prompts in three groups, each run once on the same model, with every result shown below:

• Family: four separate photos (three people and a dog) merged into one Christmas photo, including a card with the greeting written on it.
• Pets: one photo of a dog turned into six holiday portraits.
• Products: one photo of a mug turned into six holiday listing and ad shots.

Each group maps to a different way of using reference photos, and each one is a single batch rather than sixteen trips through a chat window.

How we tested#

• Made-up subjects. The three people, the dog and the mug are all AI-generated, so nobody's real photos are in this post. The people photos are casual phone-style shots, the kind you actually have on your camera roll.
• One model: Nano Banana Pro, from Google's Nano Banana family (the image models behind Gemini), as available in October 2026. If you want to see how it compares with other models on faces, we tested that in our Halloween AI photo prompts post.
• Same settings: 2K resolution, one attempt per prompt, no cherry-picking. All 16 images rendered on the first try.
• Every prompt ends with a "keep it" line that names the details people would notice first: faces for the family, markings for the dog, glaze and print for the mug.

One run per prompt is a small sample. Read what follows as tendencies, not rules.

Christmas family photo prompts#

The family scenes use Combine images: all four reference photos are sent together with every prompt, and the model puts everyone into one picture. The prompt has two parts. The scene comes first. Then a fixed line tells the model who is in the photos and that nobody may be added or dropped.

Here is the line we used. Describe each person in a few words so the model can match photo to person:

The reference images show four family members: [a woman with curly auburn hair and freckles], [a man with a short dark beard], [an older woman with short silver hair and round glasses], and [their black-and-white border collie]. Keep everyone exactly as in their photos: same faces, same features, same skin tones, same ages, same hairstyles, and the dog's exact markings. Exactly [three people and one dog], do not add or remove anyone.

And the four scenes, each followed by that line:

Christmas morning family photo: the three of them in matching red plaid pajamas sitting together on a rug in front of a decorated Christmas tree and a fireplace with stockings, the dog lying in front of them, warm tree lights, cozy candid feel, photorealistic.
Family photo on a snowy front porch at dusk: the three of them in winter coats and knit scarves standing by a front door with a big green wreath, the dog sitting in front wearing a red knit scarf, warm porch light, gentle snowfall, photorealistic.
Christmas card photo: the three of them in cream knit sweaters sitting close together on a sofa next to a Christmas tree, the dog on the sofa with them, soft warm light, the top third of the image left as a calm dark green wall with the words "Merry Christmas" in elegant white script and "Love, the Hale family" in small white letters below it, photorealistic.
Candid photo of the family decorating the Christmas tree in their living room: the older woman hanging a glass ornament, the man lifting a string of lights, the woman laughing while holding a box of ornaments, the dog watching from the floor, warm evening light, natural and unposed, photorealistic.

What came back:

• Headcount: 4 out of 4. Every image has exactly three people and one dog. Nobody was duplicated, and no extra relatives appeared in the background.
• Faces held. Cropped side by side with the original photos, all three are clearly the same people in all four scenes, including the porch shot at dusk.
• The card text was spelled right. "Merry Christmas" and "Love, the Hale family" both came out letter-perfect. The model read "dark green wall" as a flat green banner across the top. That works for a card, but don't expect a real wall.
• Clothes you don't mention come from the photos. In the tree-decorating scene (4) we didn't say what anyone wore, so the model dressed them in the exact outfits from their reference photos: green cardigan, gray henley, navy sweater. The other three scenes named the outfits and got them.
• Busy, many-action scenes get odd staging. In the same image the grandmother ended up perched on a second stepladder. It reads fine at a glance, but it's the one image a careful viewer would question.

How many people fit? Combine images sends every reference photo at once, so the limit is the model's reference count. Nano Banana Pro takes up to 4, which is exactly three people and a dog. For bigger families, GPT Image 2.5 takes up to 8 (we didn't test it on this). Each model's limit is listed on the models page.

Christmas pet photo prompts#

Pets use Per image: one photo of your pet, paired with every scene in the list. With several pets, upload all their photos and each one gets every scene.

Copy the list below, one prompt per line. In our run we also named our dog's specific markings ("off-center white blaze on the face", "one ear up and one folded"). Add your pet's two or three most recognizable details to the last sentence of each line.

The pet wearing a red Santa hat with a white pom-pom, sitting next to a decorated Christmas tree and a pile of wrapped presents, warm tree lights, photorealistic pet portrait. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.
The pet lying on a red plaid blanket in front of a crackling fireplace with knitted stockings hanging above, warm firelight, cozy and calm, photorealistic. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.
The pet in a snowy backyard wearing a red knit scarf, snowflakes falling and resting on its fur, soft winter daylight, photorealistic. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.
The pet peeking out of a large open gift box wrapped in red paper with a gold ribbon, tissue paper around it, playful expression, bright living room, photorealistic. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.
Close-up portrait of the pet wearing a felt reindeer antler headband, blurred golden fairy lights in the background, shallow depth of field, photorealistic. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.
Classic oil painting portrait of the pet in the style of an old master, wearing a small red velvet bow, framed by a holly wreath, dark green background, rich brushwork. Keep the pet exactly as in the reference photo: same breed, same coat colors and markings, same ear shape, same eye color; do not change the markings.

The dog was the easiest subject of the test:

• Markings: 6 out of 6. The off-center blaze and the folded ear survived every scene, including the reindeer antlers, where the model had to fit a headband between real ears and fake ones. The oil painting widened the white on the chest a little, which reads as painting style rather than a different dog.
• Vague rooms are filled with your room. The Santa hat (1) and gift box (4) prompts only said "next to a tree" and "bright living room", so the model reused the room from the reference photo: same bookshelves, same window, even the same dog bed on the floor. If your original photo was taken next to the laundry pile, describe the room you want.
• "Peeking out" means you see half a face. The gift box shot did what it was told, but the dog's face is half-hidden by the box. For a card, the Santa hat or antler shots work better.

Christmas product photo prompts#

The product shots also use Per image: one clean photo of the product, six scenes. This is the one that pays for itself if you sell anything. One photo, a full set of holiday listing images, social posts and an ad, with no props, no set and no reshoot.

The important part is the last sentence, the protection line. It names what a customer would notice first. Ours named the mug's shape, glaze, rim, handle and the pine tree print. Name your product's equivalents.

The product filled or styled for the season on a rustic wooden table with pine branches and a few pine cones around it, warm string lights blurred in the background, product photography. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.
Two hands in a chunky cream knit sweater holding the product by a frosty window with snow falling outside, soft morning light, lifestyle product photo. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.
Top-down flat lay: the product sitting in an open kraft paper gift box with red tissue paper, tied twine, sprigs of fir and a cinnamon stick beside it, clean light background, product photography. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.
The product on a white marble kitchen counter next to a plate of gingerbread cookies, bright clean daylight, minimal holiday styling, product photography. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.
The product on the floor under a decorated Christmas tree among wrapped presents, warm tree lights, cozy evening mood, product photography. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.
The product on a deep red studio backdrop with scattered gold and red Christmas ornaments around it, dramatic soft lighting, holiday ad photography. Keep the product from the reference image exactly as it is: same shape and proportions, same colors and finish, same printed design; do not redesign it, do not add logos or any text to the product.

In our run the first line said "the mug filled with hot cocoa and mini marshmallows". If the scene should change the product's state (filled, lit, open, switched on), say so outright.

At thumbnail size, all six look like the same mug: same shape, speckled glaze, brown rim and base, round handle, and the tree in the same spot. The hands in the window shot (2) hold the handle properly, which is where AI hands usually go wrong. Up close, there is one difference worth knowing about:

The print is redrawn, not copied. Same style, same color, same position every time, but the branches are drawn fresh each time, and some trees come out slimmer than the original. Nobody will notice on a lifestyle or ad shot. For your main listing image, where the buyer checks the print against what arrives, use your real product photo and keep the AI shots for the secondary images. Our bulk product photos guide has a checklist for reviewing every image against the original, label and print first.

Tips for your own Christmas prompts#

The first three come straight from what we saw. The others are fixes we didn't re-run, so try one image before a full batch.

• Name the outfit in every family scene. Otherwise people show up in whatever they wore in their photo.
• Describe the room. "A living room" gets you the living room from the reference photo, mess included.
• Keep the headcount sentence. "Exactly three people and one dog, do not add or remove anyone" got 4/4 here. Change the numbers, keep the sentence.
• One action per person, at most. The busier the scene, the stranger the staging. If a candid scene looks off, cut it to two actions.
• Short greetings, checked letter by letter. Two lines came out perfect. Treat anything longer as a draft.
• Name what makes your pet or product recognizable. "Same markings" works better when you also say which markings.

Run the whole set as one batch#

In a chat app, each of these is a separate round of upload, paste, wait, download: sixteen rounds for this post, more if you redo any. With a bulk image generator each group is one batch.

1. Start a New batch and paste a group of prompts into the prompt box, one prompt per line.
2. Upload your photos and choose how to use them under How to use the references:
   • Combine images for the family: every photo goes into every image.
   • Per image for pets and products: each photo gets every scene. Three pets × six scenes is 18 portraits in one run.
3. Pick the model and size. We used Nano Banana Pro at 2K, 6 credits per image. Our three batches were 4 + 6 + 6 = 16 images, 96 credits.
4. Check the total and start. The form shows the image count and credits before you spend anything. Download everything as one ZIP.

Run one image of each group first, check faces, markings and print up close, then run the rest. If you're making a Christmas book for the kids as well, our Christmas coloring page prompts work the same way.

FAQ#

Can I turn separate family photos into one Christmas photo with AI? Yes. Upload each person's photo and use a mode that sends all of them with the prompt (Combine images in BulkImagen). Describe each person in a few words and say exactly how many people and pets should appear. In our test, four separate photos became four group scenes with the right headcount every time.

How do I keep my pet looking like my pet? End the prompt with a line naming the breed, coat colors, markings, ear shape and eye color, and mention the two or three details people recognize. Our dog's off-center blaze and folded ear survived all six scenes, including an oil painting.

Can AI write "Merry Christmas" on a card correctly? Short text, usually yes. "Merry Christmas" and a one-line signature came out spelled correctly on the first try in our test. Check every letter before printing, and keep the message short.

Are AI product photos good enough for my shop listings? For lifestyle, gift and ad images, yes: our mug kept its shape, glaze and print position in all six scenes. Small printed designs are redrawn rather than copied, so use your real photo as the main listing image.

Do these prompts work in Gemini or ChatGPT? Yes. They're plain language and work with any model that accepts reference photos. We ran them on Nano Banana Pro, one of the Nano Banana models behind Gemini's image tools.