ChatGPT Christmas Photo Prompts: Holiday Portraits From Your Own Photo
ChatGPT Christmas photo prompts that turn a reference photo into a holiday portrait on GPT Image 2 and Nano Banana Pro.

Copy-ready chatgpt christmas photo prompts that turn a reference photo into a holiday portrait — fireplace, snowy street, Santa hat, and more — each shown as a real Omnely reference edit.




Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Place her in a cozy living room, seated near a lit fireplace, wearing a cream knit sweater, warm firelight on her face, a softly blurred decorated tree in the background, relaxed natural smile, portrait photograph.

Single GPT Image 2 reference edit at 1K (10 credits): cream sweater, fireplace and tree. Pose changed to chin-on-hand; identity well kept. Generated fictional person.
Paste that into GPT Image 2 on Omnely, attach one clear reference selfie, and generate — new users get free credits to try, and image models run on a paid Omnely plan.
How these chatgpt christmas photo prompts actually work
The formula behind every chatgpt christmas photo prompt below is the same: lock the face, swap the world around it. Each prompt opens with the same identity sentence ("keep her same face, facial features, skin tone, eye color, age, and dark hair"), then changes clothing, lighting, location, and framing. That one instruction is doing most of the work — without it, holiday wardrobes and warm firelight tend to drift the face into a different person.
A quick note on tooling: Omnely does not have a photo-editor UI. There is no brush, no mask, no layer panel. These are reference-image generations — you upload a source photo, the model reads it as visual context, and it writes a new photo from scratch that tries to preserve your likeness. That distinction matters: the model is not retouching your picture, it is generating a fresh one that resembles you. Expect pose changes, hair variations, and small facial softening. For the broader non-holiday version of this technique, see the parent guide on ChatGPT photo prompts.
The source photo used for every example below

Synthetic, generated fictional person. Nano Banana 2 selfie reference, 1K, 8 credits, on Omnely.
Every result in this post started from a single reference image: a generated fictional adult woman (not a real person) created as a plain phone-style selfie — dark hair in a loose bun, a dark green cable-knit sweater, gold studs, soft indoor window light, no makeup. She is a synthetic stand-in used so these prompts can be tested honestly without using anyone's real face. The selfie itself was generated on Nano Banana 2 at 1K. All Christmas edits below use her as the only reference.
Example 1 — Fireplace portrait
The fireplace prompt (shown at the top of this post) was run on GPT Image 2 at 1K, 10 credits. This is a single example, not a benchmark. The sweater, firelight falloff, and softly blurred tree all landed, and the face stayed close to the reference. The model did change her pose — chin resting on hand instead of a straight-on selfie — which is typical when the prompt describes a seated scene rather than a camera angle.
Example 2 — Snowy city street at dusk
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Place her on a snowy city street at dusk, light snow falling, shop windows glowing with warm lights, a camel wool coat and a burgundy scarf, candid standing portrait, cinematic 35mm look.

Single GPT Image 2 reference edit at 1K (10 credits): camel coat, burgundy scarf, falling snow. Identity well kept. Generated fictional person.
GPT Image 2, 1K, 10 credits. The camel coat and burgundy scarf came out clean, falling snow reads as motion rather than noise, and the warm shop-window bokeh gives it the asked-for 35mm feel. Identity held well on this one — a single run, so results will vary.
Example 3 — Solo Christmas card portrait
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Turn this into a formal Christmas card portrait of her alone: she stands slightly off-center in a deep green velvet dress, a softly lit decorated tree behind her, warm studio lighting, leave a clean empty area at the top of the frame for a greeting to be added later, no words in the image.

Single GPT Image 2 reference edit at 1K (10 credits): formal solo card-style portrait, no extra people. Empty space at the top for a greeting is limited. Generated fictional person.
GPT Image 2, 1K, 10 credits. This is a solo card portrait — the source photo is one person, so no extra family members were invented. The deep green velvet dress and formal studio lighting landed, and the model respected the "no words in the image" rule. Honest caveat: the requested empty space at the top is tighter than specified. The tree and subject fill most of the frame, so if you plan to add a greeting afterward, generate with a bit more top margin in mind or crop a wider aspect ratio before overlaying type.
Example 4 — Santa hat studio headshot
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Studio holiday head-and-shoulders portrait, she wears a simple red Santa hat and a white knit sweater, clean light-gray seamless backdrop, soft beauty-dish lighting, natural smile, no props other than the hat.

Single GPT Image 2 reference edit at 1K (10 credits): red Santa hat, white sweater, gray backdrop. Identity well kept. Generated fictional person.
GPT Image 2, 1K, 10 credits. Clean studio setup — red Santa hat, white sweater, seamless gray background, soft key light. Identity was well kept on this single example. This is a holiday-only counterpart to the cleaner studio look covered in the ChatGPT headshot prompt guide; use that one when you want a professional headshot without the hat.
Example 5 — Polaroid by the tree
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Recreate the scene as a slightly faded instant-film snapshot of her standing beside a decorated Christmas tree in a living room, warm lamp light, visible white instant-film border with no handwriting or printed date, casual candid pose.

Single GPT Image 2 reference edit at 1K (10 credits): white instant-film border with no handwriting; she is still in a green cable sweater. Identity well kept. Generated fictional person.
GPT Image 2, 1K, 10 credits. The white instant-film border appeared without any fake handwriting or printed date (the "no text" guardrail did its job). Interestingly, she is still wearing a green cable sweater similar to the reference — the prompt didn't specify an outfit, so the model held the one it saw in the source. Identity well kept.
Example 6 — Vintage 1950s Christmas
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Restyle as a vintage 1950s Christmas photograph: slightly faded color print, she wears a fitted red party dress and a string of pearls, standing by a mid-century aluminum tree with color wheel lights, soft film grain, period living-room interior.

Single GPT Image 2 reference edit at 1K (10 credits): red dress, pearls, aluminum-style tree. Hair is worn down rather than in the original bun; face otherwise close. Generated fictional person.
GPT Image 2, 1K, 10 credits. The period styling came through — fitted red dress, pearls, aluminum tree, faded print look, subtle grain. Honest note: her hair is softer and down here rather than in the reference bun. The face is otherwise close, but this is a good reminder that period-specific prompts ("1950s") will often pull the whole look — including hair — toward the era. Identity mostly kept.
Example 7 — Wreath on the front door
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Place her on the front step of a house, standing beside a dark wooden door with a large evergreen wreath and a warm porch light, she wears a charcoal wool coat and red gloves, gentle evening snow, photorealistic holiday portrait.

Single Nano Banana Pro reference edit at 1K (15 credits): charcoal coat, red gloves, wreath and snow. Hair is down; face is close but slightly smoother than the selfie. Generated fictional person.
This one was run on Nano Banana Pro at 1K, 15 credits, as a side-by-side test. The wreath, charcoal coat, red gloves, and porch-light glow all landed cleanly. Her hair is down rather than in the bun, and the face is slightly smoother than the reference — a small amount of skin softening showed up. A single example, not a benchmark against GPT Image 2.
Example 8 — Window seat with cocoa
Edit the uploaded reference photo of this generated fictional adult woman. Keep her same face, facial features, skin tone, eye color, age, and dark hair. Do not add other people or change her identity. Photorealistic, no text, no logos, no watermark. Place her in a window seat at night, holding a mug of hot cocoa, frost on the window glass, a plaid blanket over her knees, a softly blurred tree with warm lights behind her, candid relaxed portrait.

Single Nano Banana Pro reference edit at 1K (15 credits): frosted window, cocoa, plaid blanket. Identity well kept, bun returns. Generated fictional person.
Nano Banana Pro, 1K, 15 credits. Frost on the glass, plaid blanket, warm tree lights in the background, green knit sweater, and the bun came back on this one. Identity well kept. The scene composition — subject anchored to a window, tree lights pushed to blur — gives this the most candid holiday-portrait feel of the eight runs.
GPT Image 2 vs Nano Banana Pro on these specific runs
Two models, two different rendering personalities on this single reference. Not a benchmark — just what was observed on these eight examples.
| Trait | GPT Image 2 (examples 1–6) | Nano Banana Pro (examples 7–8) |
|---|---|---|
| 1K credit cost | 10 credits | 15 credits |
| Reference images accepted | Up to 16 | Up to 14 |
| Identity hold on these runs | Well kept across 6 of 6, with a hairstyle shift in the 1950s edit | Close but slightly smoother skin on the wreath run |
| Prompt literalism (borders, "no text") | Followed the Polaroid border and no-text rule cleanly | Followed wardrobe and scene instructions closely |
| Where it shined here | Editorial, studio, and period looks | Atmospheric, warm candid scenes |
Both held identity on a difficult task (holiday wardrobe swaps). On a different reference photo, the ranking could flip — this is one example per cell, not a scored test.
More Christmas-only prompt templates (text only, not run)
These are additional holiday chatgpt christmas photo prompt templates to adapt. They were not generated for this post, so expect your own first run to need a tweak. They do not overlap with the general photo prompts in the parent guide.
Kitchen baking scene
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. Place the subject in a warm kitchen at golden hour, rolling out gingerbread dough on a floured wooden counter, a red apron over a cream shirt, cookie cutters and a small decorated tree visible in the background, candid three-quarter portrait.
Stocking fireplace close-up
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. The subject kneels in front of a stone fireplace mantel decorated with three knit stockings and pine garland, wearing red plaid pajamas, soft morning light from the left, warm color grade, candid portrait from a low angle.
Snowy forest walk
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. Place the subject walking on a quiet snowy forest path lined with pine trees, wearing a cream fisherman sweater, a long navy coat, and leather gloves, soft overcast winter light, breath visible in the cold air, cinematic wide portrait.
Ice-skating rink at night
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. Place the subject at an outdoor ice-skating rink at night, string lights overhead, white skates, a red beanie and a gray wool coat, soft bokeh from distant holiday lights, candid motion portrait, 50mm lens look.
Gift-wrapping flat scene
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. Overhead three-quarter angle: the subject sits on a rug surrounded by kraft-paper gifts, spools of red ribbon, and pine clippings, wearing a mustard cardigan, warm lamp light, candid relaxed portrait.
Train station holiday travel
Edit the uploaded reference photo of this generated fictional adult. Keep the same face, facial features, skin tone, eye color, age, and hair color. Do not add other people or change identity. Photorealistic, no text, no logos, no watermark. Place the subject on a vintage train station platform at dusk, snow dusting the ground, a leather weekend bag beside them, wool coat and a long striped scarf, soft tungsten lamp light, cinematic portrait with a shallow depth of field.
FAQ
Does the prompt work if the reference photo already has a Christmas background?
Yes, but it can confuse the model. The scene in the prompt (fireplace, snowy street, tree) overrides whatever is already in the reference. If the source is already holiday-themed, the output can blend both scenes in unexpected ways. A neutral, well-lit reference — plain wall, soft daylight, face clearly visible — gives the cleanest identity hold on holiday prompts.
Why does the pose change even when the face holds?
The identity lock in the prompt only constrains facial features, skin tone, age, and hair. It does not pin the pose. When the prompt says "seated near a fireplace" or "walking on a forest path," the model will build the body language that fits the scene. If a specific pose matters, describe it explicitly ("facing the camera, hands folded in lap, straight-on framing") alongside the scene.
Can the same prompt be used for a male subject or a child?
Yes — swap the pronoun and age reference ("adult man in his 30s," "boy around age 8"), and keep the identity sentence intact. The scene description, wardrobe, and lighting work the same way. Test one example first, since some wardrobe cues (velvet dress, pearls) obviously need to be rewritten to match.
Try the fireplace prompt
Open GPT Image 2 on Omnely, upload one clear reference photo, and paste the fireplace prompt from the top of this post. At 1K it runs on 10 credits per generation, so two or three attempts will dial in the pose and framing without burning through a plan. Hold the identity sentence, change only the scene and wardrobe, and the rest of this list becomes a holiday album from one selfie.