AI image prompt gallery
24 AI image prompts, sorted by the job you need done.
Five sections by job — photoshoot, portrait and avatar, stylization, Nano Banana, product and interior. Every prompt is open in full. Copy it, or open that card in PortraitBot.
01 / 05
AI photoshoot prompts
An AI photoshoot prompt is not a request for a nice photo; it is four decisions the model cannot guess for you — who is in frame, where the light comes from, what the skin and the fabric are made of, and how much of the scene the crop keeps — which is why every card below opens with a fictional adult subject, names its light source in plain words such as a window, a direct flash or low backlit sun, fixes the palette, and closes with the no-text and no-logo clause. Take the card nearest to what you pictured, copy the whole prompt, and change one axis at a time — location first, then light, then wardrobe — so you can see which edit actually moved the frame and which one only added words.
- Subject
- One fictional adult, named before anything else
- Light
- One light source per prompt, never two
Window Editorial
A calm daylight frame for an about page or a first author photo.
Prompt265 characters
Editorial portrait of the person in the reference photo — keep their face, hairstyle and proportions instantly recognizable. Beside a large window, soft directional daylight, quiet luxury wardrobe, natural skin texture, candid magazine composition, no text or logo.
- Light
- Soft directional daylight from one window
- Frame
- Candid magazine crop, subject off-centre
Run this section in PortraitBot
The bot opens this prompt in your language and does not rewrite it. Nothing is charged until you confirm the generation.
Open the featured promptNearby on VUST
Night Flash
A direct-flash night look with edge, for a social avatar.
Prompt247 characters
Night street portrait of the person in the reference photo — keep their face and hair exactly as in the photo. Direct on-camera flash, deep city shadows, glossy highlights, confident fashion pose, authentic compact-camera texture, no text or logo.
Coastal Editorial
A full-length fashion frame built on wind and sea haze.
Prompt264 characters
Coastal fashion editorial featuring the person in the reference photo — keep their face and body proportions recognizable. Wind-shaped clothing, pale cliffs and sea haze, premium natural light, elegant full-body framing, realistic skin and fabric, no text or logo.
Alpine Luxury
A winter travel portrait for a premium seasonal post.
Prompt259 characters
Luxury winter travel portrait of the person in the reference photo — keep their face and hair exactly as photographed. On an alpine terrace, refined knitwear and outerwear, crisp mountain air, warm low sunlight, premium editorial color grade, no text or logo.
More in this section
Travel Postcard
A sunlit old-town walk for a travel post.
Prompt240 characters
Dreamlike but photorealistic Mediterranean travel scene with the person in the reference photo — keep their face unchanged. Walking through a sunlit old-town street, candid composition, natural colors, premium 35mm texture, no text or logo.
Golden Hour
A golden-hour lifestyle frame in an open field.
Prompt245 characters
Natural golden-hour lifestyle portrait of the person in the reference photo — keep their face and hair exactly as photographed. In an open field, candid movement, soft backlight, realistic skin texture, subtle analog film grain, no text or logo.
Street Fashion
A walking street-fashion frame in plain daylight.
Prompt236 characters
Street-fashion editorial with the person in the reference photo — keep their face and build true to the photo. Confident walking pose, layered contemporary outfit, modern architecture, crisp daylight and natural motion, no text or logo.
02 / 05
AI portrait and profile picture prompts
Prompts for a portrait or a profile picture solve one problem: the face has to stay recognisable after the light, the background and the wardrobe have all been styled, which is why every card below fixes the light source and the backdrop first, the clothing second, and only then adds the stylistic detail a model can push without damaging anatomy. If the output is going to be an avatar, add one framing requirement to the prompt — head and shoulders centred, with margin at the edges — and never name two light sources in one prompt, because mixing a softbox with street neon is what usually breaks a face and turns a portrait into a collage.
- Frame
- Head and shoulders centred, margin at the edges
- Light
- Style the background, never the anatomy
Film Noir
A high-contrast black-and-white portrait for a dramatic profile.
Prompt259 characters
Cinematic black-and-white portrait of the person in the reference photo — keep their facial features and hair unchanged. Sculpted hard light through blinds, deep shadows, classic noir wardrobe, subtle film grain, dramatic but natural anatomy, no text or logo.
- Light
- Hard light through blinds
- Palette
- Black and white with fine film grain
Run this section in PortraitBot
PortraitBot also builds a portrait from your own selfie, with the same confirm step before anything is charged.
Open the featured promptNearby on VUST
Magazine Editorial
A studio fashion frame carried by one graphic shadow.
Prompt233 characters
High-fashion studio editorial with the person in the reference photo — keep their face, skin texture and proportions true to the photo. Sculptural wardrobe, graphic shadow, clean cream backdrop, bbold negative space, no text or logo.
Cinematic Rain
A rain portrait with eye contact, warm and cinematic.
Prompt243 characters
Cinematic portrait of the person in the reference photo under gentle rain — keep their face and hairline recognizable. Warm practical lights behind, realistic wet hair and fabric, shallow depth of field, emotional eye contact, no text or logo.
Monochrome Studio
A minimal monochrome studio portrait for a work profile.
Prompt257 characters
Minimal monochrome studio portrait of the person in the reference photo — keep their facial structure and hair unchanged. Charcoal wardrobe and backdrop, precise softbox lighting, elegant geometry, realistic texture and restrained contrast, no text or logo.
More in this section
Y2K Digital
An early-digital-camera avatar with a playful gloss.
Prompt241 characters
Playful early-digital-camera fashion portrait of the person in the reference photo — keep their real face and anatomy under the styling. Glossy metallic accessories, direct flash, clean pale backdrop, controlled Y2K styling, no text or logo.
03 / 05
Anime, illustration and 3D style prompts
Anime, illustration and 3D prompts differ from photoreal ones in a single way: the technique is as much a part of the brief as the subject, and it has to be named in words that describe the material rather than an artist or a studio. That is why the cards below say visible paper texture and confident brushwork instead of a painter's name, crisp line art and a layered environment instead of a series title, and tactile clay surfaces instead of a toy brand — a description like that is reproducible, leans on nobody's signature, and leaves you one free axis, the character, which you can change without dismantling the style.
- Material
- Name the technique as a material, not as an author
- Subject
- Keep one free axis: the character
Watercolor Story
A watercolour illustration for a soft editorial header.
Prompt215 characters
Expressive watercolor illustration of a fictional adult traveler, visible paper texture, confident brushwork, soft atmospheric background, coherent anatomy and silhouette, refined editorial palette, no text or logo.
- Material
- Visible paper texture, confident brushwork
- Excluded
- No text, no logo
Run this section in PortraitBot
A stylised square also works as a real chat sticker once it is sized for the sticker format.
Open the featured promptNearby on VUST
Clay 3D
A clay-style 3D character for a friendly avatar.
Prompt270 characters
Premium clay-style 3D collectible figure of the person in the reference photo — face instantly recognizable (features, hairstyle, proportions) in tactile clay. Clean studio backdrop, soft product lighting. Replace the selfie room entirely, not a filter. No text or logo.
Anime Cel
A cel-anime portrait with clean line art.
Prompt268 characters
Polished cel-anime portrait of the person in the reference photo — character stays recognizable: same face shape, hairstyle, hair and eye color. Crisp line art, expressive eyes, cinematic lighting, layered anime scene. Full illustration, not a filter. No text or logo.
Pet Sticker
A sticker-style pet portrait with a bold outline.
Prompt191 characters
Cute high-detail sticker-style portrait of a fictional corgi, bold white outline, joyful expression, clean coral backdrop, crisp 3D illustration, centered square composition, no text or logo.
More in this section
Fantasy Royal
A costume fantasy portrait that keeps hands and face real.
Prompt244 characters
Cinematic fantasy portrait of the person in the reference photo as a ruler — keep their face and hands recognizable, do not idealize the features. Refined embroidered costume, atmospheric palace interior, premium film lighting, no text or logo.
04 / 05
Nano Banana prompts
Nano Banana is the common name for Google's Gemini image model, and it is the image model PortraitBot currently runs, so the prompts in this gallery go into it without a rewrite. The three cards below were picked for what that class of model is asked for most and handles best — a composite scene in which reflections, a period language or a change of scale all have to agree inside one frame instead of falling apart into separately pretty pieces — and the gallery still promises no fixed model, because a prompt that describes a scene rather than the settings of one engine keeps working when the engine changes.
- Subject
- One composite scene, not a list of nice objects
- Excluded
- Describe the scene, never the engine settings
Cyber Neon
A rain-lit neon scene where every reflection has to agree.
Prompt250 characters
Stylized cinematic portrait of the person in the reference photo — keep their real face and anatomy intact under the styling. Rain-lit neon street, cyan and coral reflections, sharp fashion photography, controlled futuristic details, no text or logo.
- Light
- Cyan and coral neon in a wet street
- Subject
- A real face inside a stylised scene
Run this section in PortraitBot
The bot names the model that actually returned your image in the delivery caption.
Open the featured promptNearby on VUST
Retro Future
A 1970s space-age scene that must hold one period language.
Prompt236 characters
Retro-futurist editorial scene with the person in the reference photo — keep their face and hands realistic and recognizable. Optimistic 1970s space-age forms, warm cream and coral palette, cinematic practical lighting, no text or logo.
Miniature World
A miniature diorama whose scale trick must read at once.
Prompt172 characters
Whimsical miniature diorama of tiny fictional chefs building an oversized pastry, macro photography, realistic materials, warm bakery light, playful scale, no text or logo.
05 / 05
Product, interior and collectible prompts
Product, interior and collectible prompts differ from portraits because the subject cannot pose: material, reflections and the empty space around the object decide the picture instead, and any vagueness in those three words shows up immediately. So the cards below name the material literally — refraction and water ripples, natural oak and stone, a molded display tray — keep generous space around the object for a caption or a price you will add later, and end with the no-readable-text and no-logo clause, because generated lettering on packaging almost always breaks and someone else's logo makes the image unusable on a listing.
- Material
- Name the material literally, not by mood
- Frame
- Leave space for the caption you will add later
Collector Figure
A collectible-figure product shot for a packaging mock-up.
Prompt267 characters
Premium collectible action-figure PRODUCT SHOT of the person in the reference photo, rendered as a physical PVC doll figure, not a live photo. Vacuum-formed blister pack, molded display tray, studio lighting. Forbidden: selfie-room overlay, no readable text or logos.
- Material
- Molded display tray, detailed accessories
- Excluded
- No readable text, logo or brand
Run this section in PortraitBot
The same prompt shape works for a listing hero once you swap the object description.
Open the featured promptMore examples in the channelNearby on VUST
Hero Card
A collectible-card portrait with graphic depth.
Prompt276 characters
Cinematic collectible-card portrait of the person in the reference photo as an athlete — face and build recognizable. Dynamic rim light, layered card graphics, foil-like accents, designed card art, not a photo with a border overlay. No readable text, numbers, logos or brands.
Glass Product Hero
A glass product hero for a listing or an ad.
Prompt206 characters
Premium product hero shot of an unbranded translucent glass fragrance bottle, refracted coral light, water ripples, clean luxury composition, realistic reflections, generous negative space, no text or logo.
Room Makeover
An interior concept with a livable layout and no people.
Prompt197 characters
Photorealistic warm-modern living room concept, sculptural furniture, natural oak and stone, soft afternoon light, practical livable layout, editorial interior photography, no people, text or logo.
Five sections answer five jobs, and every one of the 24 cards pairs a finished image with the full prompt that made it — in English here, in Russian on vust.io — printed in full, never hidden behind a reveal. PortraitBot runs the same prompt you read on this page.
Each card is a reusable generation protocol: subject, light, materials, composition, and the no-text/no-logo clause. Copy the prompt or open it in PortraitBot — the bot keeps it word for word until you confirm the generation.
Browsing is free
All 24 images and prompts are visible here without using Portrait USD Balance. Opening or browsing the weekly inspiration card is also free. Copying a prompt or opening a card in PortraitBot does not use balance. If you generate, balance is used only after a generated image is successfully delivered. A Portrait Pass is not required.
Process
How to use the gallery
- 01
Pick the section that matches your job, then read its answer first.
- 02
Browse the 24 visual references and read each prompt in full — nothing is hidden.
- 03
Copy the prompt from a card, or open that card in PortraitBot.
- 04
The bot opens this prompt in your language and does not rewrite or enhance it.
- 05
Review the ordinary generation confirm before asking for an image.
What is included
- 24 original images made from these exact prompts
- 24 prompts in this page’s language, each printed in full
- Five sections by job, each opening with its answer
- Fictional subjects or unbranded objects
- Free browsing, copy, and per-card bot links
What is not implied
- The gallery does not promise a fixed model
- A reference is not a guarantee of identical generation
- Trying a prompt is an ordinary generation action, not free gallery browsing
- No third-party logos or readable branded text are intended
Examples
Three prompt-editing habits
Keep composition and lighting, but swap the palette or setting first.
Retain the no-text and no-logo clauses when you want a clean visual.
Describe a fictional adult or unbranded object instead of asking to copy a real identity.
Prompt anatomy
How to write an image prompt
A working image prompt is five decisions in one sentence: who or what is in frame, where the light comes from, what the surfaces are made of, how much of the scene the crop keeps, and what must not appear at all. The fifth is the one people skip, and it is the one that removes stray lettering, watermarks and other people's logos — which is why the last clause of every prompt in this gallery forbids text and logos. Change one axis at a time and hold the other four still: that is how you learn which edit moved the frame and which one only made the sentence longer.
- Subject
- Who or what is in frame, and that it is fictional
- Light
- One named light source
- Material
- What the surfaces are made of
- Frame
- How much of the scene the crop keeps
- Excluded
- What must not appear at all
FAQ
AI image prompt questions
- Are the gallery images owned by VUST?
- Yes. All 24 images were created by VUST from these exact prompts, for this gallery and for PortraitBot’s inspiration collection.
- Which language are the prompts in?
- The language of the page you are reading: English on vust.ai, Russian on vust.io. Both versions are written by hand, not machine translations of each other, and copying gives you the exact text shown on this page.
- Which model do these prompts target?
- None in particular. PortraitBot currently generates with Google's Gemini image model, the one commonly called Nano Banana, but each prompt describes a scene rather than the settings of one engine, so it survives a change of model.
- Does browsing use Portrait USD Balance?
- No. Browsing this page and the inspiration cards is free. An ordinary Create action uses balance only after successful image delivery.
- Can I expect an identical image from the prompt?
- No. Each card is a visual reference and a starting prompt, not a guarantee of an identical result.