Gemini photo prompts.
40 copy-paste prompts for portraits of men, women and couples, saree and traditional looks, professional headshots and 3D figurines. Plus the one line that stops Gemini quietly changing your face.
GPTPrompts.AI Editorial
Prompts structured for image models that edit from a reference photo · Last updated August 2026
Why most Gemini photo prompts disappoint
Almost every prompt list online is a wardrobe description: put me in a suit, put me on a beach. That produces a generic image with a face that is nearly you.
The prompts here add three things those lists leave out. An explicit instruction to preserve your likeness. A named lens and light direction, which is what photographers actually control. And a ban on skin smoothing, which is the single clearest tell that an image was generated.
Ultra realistic portraits, the base technique
Start here. Every other group on this page is this technique with a different wardrobe and setting. The parts that matter are the lens, the light direction and the instruction to keep natural skin texture.
Using the uploaded photo as reference, keep the facial features, skin tone and hairstyle exactly as they are. Create a photorealistic portrait: 85mm lens at f/2.0, soft window light from the left, gentle falloff into shadow on the right, plain warm grey background slightly out of focus. Natural skin texture with visible pores, no smoothing or beauty filter, catchlights in the eyes. 4:5.
From the uploaded photo, preserve the face precisely. Photorealistic environmental portrait: the person standing near a large window in a quiet room, 35mm lens, natural overcast daylight, muted colour palette, relaxed posture, looking slightly away from camera. Fine grain, no retouching, documentary feel. 4:5.
Using the uploaded photo, keep the likeness identical. Cinematic portrait with a single warm key light from the upper right and deep shadow on the opposite side, dark background, 85mm, shallow depth of field, serious and calm expression, film grain. Photorealistic, no text. 9:16.
From the uploaded photo, face unchanged. Golden hour outdoor portrait, low sun behind the subject creating a rim light on the hair, warm haze, blurred greenery in the background, 135mm compression, natural skin with slight shine. Photorealistic. 4:5.
Using the uploaded photo and preserving the face exactly, create a high-key studio portrait: white seamless background, large soft light source directly in front and slightly above, minimal shadow, clean and modern, sharp focus on the eyes, natural skin texture retained. 1:1.
Prompts for men
Grooming detail and fabric texture are what separate a convincing male portrait from a mannequin. Name the fabric, not just the garment.
Using the uploaded photo, keep the face, skin tone and hairstyle exactly as they are. The man wears a charcoal wool overcoat over a fine-knit black roll neck. Standing on a city street at dusk, blurred lights behind, cool blue ambient light with a warm rim from a shop window. 85mm, photorealistic, natural skin texture, confident but relaxed. 4:5.
From the uploaded photo, face preserved precisely. The man wears a crisp white linen shirt with the top button open, sleeves rolled. Seated at an outdoor cafe in soft morning light, warm tones, shallow depth of field, natural stubble texture visible, candid expression mid-thought. Photorealistic. 4:5.
Using the uploaded photo, keep the likeness identical. Rugged outdoor portrait: the man in a weathered olive field jacket, overcast mountain light, wind in the hair, muted colours, slight moisture on the skin, 50mm lens. Photorealistic documentary style, no text. 9:16.
From the uploaded photo, preserve the face. The man wears a well-fitted navy suit with a white shirt and no tie, standing against a dark textured wall, single soft light from the left, hands in pockets, calm expression. Editorial fashion photography, natural skin, 85mm. 4:5.
Using the uploaded photo and keeping the face unchanged, create a monochrome portrait: the man in a plain dark t-shirt, dramatic side lighting, deep blacks, high contrast, visible skin texture and stubble, looking directly at the camera. Black and white, photorealistic, 85mm. 1:1.
Professional headshots for men and women
For LinkedIn and company pages. The brief is deliberately boring: approachable, current, and obviously the same person who turns up to the meeting.
Using the uploaded photo, keep the facial features, skin tone and hairstyle exactly as in the original. Corporate headshot: mid-chest crop, facing camera, slight natural smile, wearing a well-fitted navy blazer over a plain shirt. Soft even light from the front left, plain light grey background with gentle gradient, 85mm at f/2.8. Natural skin texture, no beauty filter. 1:1.
From the uploaded photo, preserve the face precisely. Modern office headshot: the person standing in a bright workspace, blurred glass and greenery behind, natural daylight, smart casual dress, relaxed confident posture, warm and approachable expression. 85mm, shallow depth of field, photorealistic. 4:5.
Using the uploaded photo with the face unchanged, create a clean editorial headshot on a plain white background, soft frontal lighting with a subtle shadow under the chin, wearing a simple dark top, direct eye contact, neutral confident expression. Natural skin retained, no retouching. 1:1.
From the uploaded photo, face identical. Outdoor professional portrait: standing in soft open shade, blurred city background, smart jacket, natural daylight with no harsh shadow, friendly expression, 85mm. Photorealistic, suitable for a company website. 4:5.
Prompts for women
The most common failure here is over-smoothing, which the model will do unless you forbid it. Every prompt below asks for natural skin explicitly.
Using the uploaded photo, keep the facial features, skin tone and hairstyle exactly as they are. The woman wears a flowing cream dress, standing in a sunlit field at golden hour, warm backlight through the hair, soft haze, 85mm at f/1.8. Natural skin texture with no smoothing, gentle expression looking slightly off camera. Photorealistic. 4:5.
From the uploaded photo, face preserved precisely. Editorial portrait: the woman in a structured black blazer against a deep burgundy backdrop, single soft light from the right, strong shadow on the left, direct confident gaze. Natural skin texture, 85mm, fashion photography lighting. 4:5.
Using the uploaded photo and keeping the likeness identical, create a candid indoor portrait: the woman by a window on an overcast day, soft diffused light, neutral knitwear, hair loose, mid-laugh, warm and unposed. Fine grain, natural skin, 50mm. Photorealistic. 4:5.
From the uploaded photo, face unchanged. Evening portrait: the woman in an elegant deep green outfit, city lights blurred behind, warm practical lighting from a nearby source, shallow depth of field, calm expression. Natural skin texture retained, 85mm, photorealistic. 9:16.
Using the uploaded photo, preserve the face. Minimal beauty portrait: plain soft grey background, large diffused light directly in front, hair pulled back simply, minimal makeup, close crop from shoulders up, sharp focus on the eyes. Real skin texture with pores visible, no airbrushing. 1:1.
Saree and traditional wear
Naming the fabric and the drape is what gets a real saree rather than a vague wrap. Silk, cotton and chiffon fall completely differently and the model knows it.
Using the uploaded photo, keep the facial features, skin tone and hairstyle exactly as they are. The woman wears a deep red Kanjivaram silk saree with a gold zari border, draped in classic Nivi style, traditional gold jewellery, hair in a low bun with jasmine. Standing in a warmly lit heritage courtyard, soft evening light, 85mm. Photorealistic, natural skin texture, no text. 4:5.
From the uploaded photo, face preserved precisely. The woman wears a pastel chiffon saree with a delicate embroidered border, light and flowing, soft breeze catching the pallu. Outdoors in open shade, gentle daylight, blurred greenery behind, minimal jewellery. Photorealistic, natural skin, 85mm. 4:5.
Using the uploaded photo with the face unchanged, the woman wears a handloom cotton saree in indigo with a simple white border, minimal jewellery, hair tied back. Seated indoors near a window with soft daylight falling across her, quiet and warm mood, documentary style. Photorealistic. 4:5.
From the uploaded photo, keep the likeness identical. Bridal portrait: the woman in a richly embroidered red and gold silk saree with traditional bridal jewellery and a maang tikka, warm candlelit interior, shallow depth of field, soft glow. Photorealistic, natural skin texture, no text. 9:16.
Using the uploaded photo and preserving the face, the man wears a cream silk kurta with a churidar and a contrasting Nehru jacket in deep maroon. Standing in a softly lit traditional interior, warm side light, calm confident posture, 85mm. Photorealistic, natural fabric texture. 4:5.
Couple photos
The hardest category, because models blend features between adjacent faces. Name each person by position and state that both faces must hold.
Using the uploaded photo of two people, keep both faces exactly as they are: the person on the left and the person on the right must each retain their own facial features, skin tone and hairstyle with no blending between them. Photorealistic portrait of the couple standing close, soft golden hour backlight, blurred open field behind, coordinated neutral clothing, relaxed and natural. 85mm at f/2.0. 4:5.
From the uploaded couple photo, preserve both faces precisely and separately. Candid indoor scene: the two seated together by a window on an overcast day, soft daylight, mid-conversation, one laughing, warm neutral tones, natural skin texture. 50mm, documentary style, photorealistic. 4:5.
Using the uploaded photo of the couple and keeping each face identical to the original, create an evening portrait: the two standing on a city rooftop, blurred lights behind, warm practical lighting, smart eveningwear, quiet and intimate mood, shallow depth of field. Photorealistic, 85mm. 9:16.
From the uploaded couple photo, both faces unchanged and distinct. Traditional portrait: the woman in a deep red silk saree with gold jewellery, the man in a cream kurta with a maroon Nehru jacket, standing together in a warmly lit heritage courtyard, soft evening light. Photorealistic, natural skin, no text. 4:5.
Using the uploaded photo of two people, keep both likenesses exactly. Walking shot from behind and slightly to the side, hands held, on a quiet tree-lined path in soft morning light, gentle motion, warm muted colours, faces visible in three-quarter profile. Photorealistic, 50mm. 16:9.
3D figurines and model renders
The viral one. Describing a physical object with a material, a base and product lighting is what makes this work, rather than asking for a 3D version.
Using the uploaded photo, turn the person into a detailed collectible figure standing on a round display base. PVC material with a matte finish and subtle gloss on the eyes, high detail on fabric folds and hair, studio product lighting, neutral grey background. Keep the face clearly recognisable. Photorealistic product photography. 1:1.
From the uploaded photo, create a boxed collectible figurine: the figure inside retail packaging with a clear plastic window, bold graphic panel, photographed straight on under even light. Keep the likeness recognisable, styled as a toy but faithful to the person. No text on the packaging. 4:5.
Using the uploaded photo, generate a stylised vinyl toy version of the person with slightly exaggerated proportions, large head and small body, keeping the outfit and hairstyle recognisable. Plain pastel background, soft drop shadow, clean product lighting. 1:1.
From the uploaded photo, create a 1/7 scale collectible figure posed naturally on a transparent acrylic base, photographed on a desk. In the soft-focus background, a computer screen showing 3D modelling software with the same figure open in it. Photorealistic. 4:5.
Six rules that decide the result
The preservation line does most of the work
Keep the facial features, skin tone and hairstyle exactly as in the uploaded photo, change only clothing, lighting and background. Paste that into any prompt here that you have modified, because it is the line that stops a portrait becoming a stranger.
Forbid smoothing explicitly
Ask for natural skin texture with visible pores and no beauty filter. Over-smoothed skin is the fastest way for an image to read as AI-generated, and the model will smooth by default.
Name a lens and an aperture
85mm at f/2.0 for portraits, 50mm for a natural documentary feel, 135mm for compressed background blur, 35mm when you want the room in shot. This is more effective than the word cinematic.
Describe light by direction and quality
Soft window light from the left, hard midday sun overhead, warm rim light from behind. Direction plus quality beats naming a mood, because the model can act on it.
One person per image where you can
Every additional face is another chance for the model to blend features. If a group shot keeps failing, generate people separately rather than fighting a single crowded prompt.
Refine in the same chat, do not restart
Reply with the one change you want: warmer light, less saturated, turn slightly left. Gemini keeps the rest of the context, so a short correction beats rewriting the whole prompt.
Frequently asked questions
How do I write a Gemini prompt that keeps my face?
Add an explicit preservation line. The most reliable wording is: keep the facial features, skin tone and hairstyle exactly as in the uploaded photo, change only the clothing, lighting and background. Gemini's image model is good at holding a likeness, but only if you tell it that the likeness is the part that must not move. Without that line it treats your face as one more thing it is free to restyle.
What makes a Gemini photo prompt ultra realistic?
Four things, and none of them is the word realistic. Name the lens and aperture, for example 85mm at f/2.0. Describe the light source and its direction. Ask for natural skin texture with no smoothing, which is the single biggest giveaway when it is missing. And add small imperfections: slight asymmetry, subtle grain, a stray hair. Perfectly even skin and perfect symmetry are what make an image read as AI.
Why do couple photos come out wrong more often?
Two faces means twice the chance of one drifting, and models often blend features between people standing close together. Use a source photo where both faces are clearly visible and separated, name each person in the prompt, for example the person on the left and the person on the right, and state that both faces must stay exactly as uploaded. Generating a few and picking the best works better than refining one endlessly.
Is there a Gemini prompt generator?
Yes, we run a free one that builds the prompt for you from a description of what you want, and it needs no login. It is useful when you know the result you are after but not how to specify the lens, light and framing that get you there. For photo work, generating a base prompt and then adding your own preservation line is a good workflow.
How do I make a 3D model or figurine from a photo in Gemini?
Describe a physical object rather than asking for 3D. Try: turn the person in this photo into a detailed collectible figure on a round display base, PVC material with a matte finish and subtle gloss on the eyes, studio product lighting, neutral grey background, keeping the face clearly recognisable. Naming the material, the scale, the base and the lighting is what produces a figurine rather than a flat, plasticky render.
Can I use Gemini photo prompts on the free plan?
Yes. Image generation and editing are available in the Gemini app at gemini.google.com without a paid plan, subject to usage limits. Paid tiers raise those limits rather than unlocking a fundamentally different image model, so the prompts on this page work the same either way.
What aspect ratio should I ask for?
Say it explicitly or you will get a square. Use 9:16 for stories and reels, 4:5 for an Instagram feed post, 1:1 for a profile picture, and 16:9 for a banner or a laptop background. Adding the ratio to the end of any prompt on this page is enough.
Why does Gemini add strange text to my image?
Image models are unreliable at rendering letters, so signs, badges and packaging often come out garbled. Add this to any prompt: do not render any text, letters or numbers anywhere in the image. Then add wording yourself afterwards in a normal editor if you need it.