Prompt Library

ChatGPT prompts for pictures of yourself from a reference

30 copy-paste prompts

Attach a clear photo of you, then paste a job: studio headshots, casual days, travel frames, outfit tests, yearbook looks and creative stills. Not a dating-app set.

In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 5 prompts are free instantly, no signup needed. Drafted with AI writing assistance and checked against our editorial policy.

Louis Corneloup
By Louis Corneloup · Founder, Techpresso
Last updated ·Drafted with AI assistance (how we make prompts)

Studio headshots

5 prompts

Neutral gray three-quarter

1/30

✨ What it does

ChatGPT builds a three-quarter studio picture of you on a named seamless from your reference. Keep the 4:5 if your hairline and glasses still match the upload.

Using [YOUR PHOTO] as the only face reference, make a three-quarter studio picture of me on a [BACKDROP COLOR] seamless. I wear [WARDROBE]. Soft key from camera left, white bounce on the right, catchlights in both eyes. 85mm at f/4, 4:5 crop from mid-chest. Keep my real face: same age, same hairline, same glasses if I wear them. No beauty slim, no celebrity lookalike. I am the only person in the frame.

💡

Pro tip: A front-facing source with even light locks identity better than a heavily filtered selfie.

Window-lit office crop

2/30

✨ What it does

ChatGPT places you at an office window with a named city view in the bokeh. Keep the 3:4 if your jaw and skin still match the reference.

From [YOUR PHOTO], generate a picture of me seated at a desk near a tall office window, [CITY VIEW] softly out of focus. I wear [WARDROBE]. 50mm at f/2.8, daylight from the window as the only key, a weak fill from the room. Crop waist-up, 3:4. Preserve my bone structure and skin texture; do not replace my jaw with a stock professional. Empty office, no coworkers over my shoulder.

💡

Pro tip: Name a generic view (river, rooftops) rather than a famous skyline with celebrity billboards.

Black and white Rembrandt

3/30

✨ What it does

ChatGPT makes a black-and-white Rembrandt head-and-shoulders of you on charcoal paper. Keep the 4:5 if pores and lines still belong to you.

Turn [YOUR PHOTO] into a monochrome picture of me, head and shoulders, Rembrandt key from the right, charcoal paper behind. 105mm at f/2.8, 4:5. I wear [WARDROBE]. Keep pores, stubble or laugh lines from the source; skip a fashion-smooth pass. Serious but not angry. I am the only subject. No extra silhouettes in the dark.

💡

Pro tip: Say 'retain skin texture' if the first take looks like a wax figure.

Color gel on seamless

4/30

✨ What it does

ChatGPT lights you with a named gel on seamless for a square portfolio still. Keep it if your eyes and mouth still match the reference under the color.

From [YOUR PHOTO], make a creative studio picture of me on a [GEL COLOR] gel-lit seamless, one hard key from the left and a complementary rim. 85mm at f/5.6 so both eyes stay sharp. I wear [WARDROBE]. Hold my identity under the color: same eyes, same mouth, same hair. This is a portfolio still, not a nightlife flyer. I stand alone. Square 1:1.

💡

Pro tip: Pick one gel color. Two competing gels often smear facial identity.

Bookshelf environmental

5/30

✨ What it does

ChatGPT stands you beside a named-subject bookshelf with mixed lamp and window light. Keep the 3:2 if height and face still match, and titles stay unreadable.

Generate a picture of me from [YOUR PHOTO] standing beside a [BOOK SUBJECT] bookshelf, 35mm at f/2.8, tungsten lamp plus a window. I wear [WARDROBE], one hand on a shelf. 3:2. Match my height and face to the reference; do not lengthen me into a catalog model. No other readers in the aisle. Titles stay unreadable.

💡

Pro tip: Unreadable spines avoid invented book titles on the shelf.

Casual everyday

5 prompts

Kitchen coffee morning

6/30

✨ What it does

ChatGPT makes a candid kitchen picture of you with a named mug and morning window light. Keep the 4:5 if the face still looks like your real morning, not a glamour pass.

From [YOUR PHOTO], make a candid picture of me in my kitchen holding a [MUG COLOR] mug, morning window at camera left. 35mm at f/2.2, crumbs on the counter allowed. I wear [WARDROBE]. Keep my real morning face, no makeup glamour. I am alone in the room. 4:5 for a personal album, not a dating card.

💡

Pro tip: A source photo without heavy filters matches this kitchen mood better.

Neighborhood sidewalk walk

7/30

✨ What it does

ChatGPT shows you mid-stride on a named neighborhood sidewalk in late-day sun. Keep the 3:4 if your gait and face still match, with no readable strangers.

Using [YOUR PHOTO], generate a walking picture of me on a [NEIGHBORHOOD] sidewalk, mid-stride, 50mm at f/2.0, late-day sun. I wear [WARDROBE], a tote on one shoulder. 3:4. Keep my gait and face from the reference; do not turn me into a street-style model. Background people stay faceless or absent. Not a dating-app stroll.

💡

Pro tip: Name the neighborhood type (brick row, tree lawn) instead of a famous shopping street.

Sofa reading still

8/30

✨ What it does

ChatGPT sits you on a sofa with a named paperback and throw under lamp light. Keep the 16:9 if your sitting posture and face still match the reference.

From [YOUR PHOTO], sit me on a sofa with a [BOOK GENRE] paperback, a [THROW COLOR] throw on my knees. 50mm at f/2.0, lamp at camera right. I wear [WARDROBE]. Preserve my face and how I actually sit. Empty room besides me. 16:9, a quiet home still, not a posed flirt.

💡

Pro tip: Name the genre, not a real book title, so the cover stays generic.

Bike by a canal

9/30

✨ What it does

ChatGPT pauses you with a bicycle beside a named canal under overcast light. Keep the 3:2 if body and face were not rewritten fitter.

Generate a picture of me from [YOUR PHOTO] paused with a bicycle beside a [WATER TYPE] canal, 35mm at f/4, overcast. I wear [WARDROBE], helmet in hand not on a stranger. Keep my face and body as uploaded; skip a fitness rewrite. I am the only readable person. 3:2.

💡

Pro tip: Say 'helmet in my hand' so the model does not invent a riding companion.

Stoop with a market bag

10/30

✨ What it does

ChatGPT sits you on a stoop with a named market bag and door color in open shade. Keep the 4:5 if elbows and face still match, with empty neighboring stoops.

From [YOUR PHOTO], sit me on a front stoop with a [BAG TYPE] market bag, [HOUSE COLOR] door behind. 35mm at f/2.8, open shade. I wear [WARDROBE]. Lock identity to the source, including how I rest my elbows. No neighbors on adjacent stoops with faces. 4:5 everyday still.

💡

Pro tip: Open shade sources (porch, cloudy day) composite cleaner than noon flash.

Travel frames

5 prompts

Train window seat

11/30

✨ What it does

ChatGPT seats you at a train window with a named landscape outside and an empty adjacent seat. Keep the 3:2 if any reflection still looks like you.

Using [YOUR PHOTO], make a picture of me in a train window seat, [LANDSCAPE TYPE] moving outside, 35mm at f/2.8. I wear [WARDROBE], a tote on the adjacent empty seat. Keep my real face in the glass reflection only if it matches me. No other passengers with readable faces. 3:2 travel still.

💡

Pro tip: Name the landscape (fields, coast) rather than a famous route map.

Museum stair landing

12/30

✨ What it does

ChatGPT places you on a museum stair with named stone and a skylight, looking at an unreadable label. Keep the 4:5 if you still look like yourself, not a fashion ad.

From [YOUR PHOTO], place me on a museum stair landing with [STONE COLOR] stone and a skylight above. 24mm at f/4, available light. I wear [WARDROBE], looking at an unreadable wall label. Keep my face; do not turn me into a fashion editorial. Empty stairs, no tour group. 4:5.

💡

Pro tip: Ask for unreadable labels so ChatGPT does not invent exhibit names.

Harbor railing pause

13/30

✨ What it does

ChatGPT stands you at a harbor railing with a named boat type in the bokeh. Keep the 4:5 if hair and face still match, with no extra people at the rail.

Generate a picture of me from [YOUR PHOTO] at a harbor railing, [BOAT TYPE] out of focus, 85mm at f/2.8, wind lifting the jacket hem. I wear [WARDROBE]. Hold my face and hair from the reference, including flyaways; skip a magazine glow on skin. I stand alone at the rail, no tourist extras. 4:5 travel still.

💡

Pro tip: A wind note on the jacket keeps the still from looking like a studio fake.

Trail switchback rest

14/30

✨ What it does

ChatGPT shows you resting on a trail switchback among named trees with a pack at your feet. Keep the 3:2 if posture stayed realistic and the trail ahead is empty.

From [YOUR PHOTO], show me resting on a trail switchback among [TREE TYPE], pack at my feet, one boot on a root. 35mm at f/5.6, bright overcast, no golden-hour filter. I wear [WARDROBE]. Keep my real face and a realistic hiking posture, not an athlete campaign. Empty trail ahead. 3:2 landscape.

💡

Pro tip: Name the trees (pine, birch) so the trail does not become a generic alpine ad.

Hotel lobby armchair

15/30

✨ What it does

ChatGPT sits you in a hotel lobby armchair with a named material behind and mixed practical lights. Keep the 4:5 if it still reads as you, not a luxury ad.

Using [YOUR PHOTO], sit me in a hotel lobby armchair, [LOBBY MATERIAL] behind, 50mm at f/2.2, mixed practicals. I wear [WARDROBE], a small bag on the floor. Keep my identity; this is a travel record, not a luxury campaign. No staff or guests with faces. 4:5.

💡

Pro tip: Name the material (marble, wood panel) instead of a famous hotel brand.

Outfit tests

5 prompts

Same pose, two outfits

16/30

✨ What it does

ChatGPT splits one stance of you into two named outfits on a shared seamless. Keep the 16:9 if both halves show the same face and light.

From [YOUR PHOTO], make a side-by-side picture of me in the same stance: left I wear [OUTFIT A], right I wear [OUTFIT B]. Shared [BACKDROP COLOR] seamless, 85mm at f/4, identical light. Keep my face identical in both halves. I am the only person. 16:9 split for a wardrobe test, not a makeover story.

💡

Pro tip: Describe fabric weight so the two sides do not share the same folds.

Full-length mirror studio

17/30

✨ What it does

ChatGPT shows you full-length in a studio mirror wearing the outfit you name. Keep the 3:4 if the reflection is your real face, not a prettier double.

Generate a full-length picture of me from [YOUR PHOTO] checking [WARDROBE] in a studio mirror, 35mm at f/5.6, even softboxes, seamless visible at the edges. The mirror reflection must match my real face, not a prettier double. Empty studio, no stylist in the glass. 3:4 outfit test.

💡

Pro tip: If the reflection drifts, rerun and say 'same face in the glass as in the room'.

Color-story sidewalk

18/30

✨ What it does

ChatGPT walks you down a sidewalk in a named color story with jacket, bag and shoes agreeing. Keep the 4:5 if legs were not lengthened for the clothes.

From [YOUR PHOTO], walk me down a sidewalk in a [COLOR STORY] outfit with jacket, bag and shoes agreeing. 50mm at f/2.2, open shade, storefronts blurred. Keep my face and proportions; do not lengthen legs for the clothes. No other stylish extras walking toward camera. 4:5 wardrobe test in public.

💡

Pro tip: Name three garment colors so the story stays tight.

Coat in a short stride

19/30

✨ What it does

ChatGPT shows you mid-stride in a named coat on a studio cyc with hard side light for fabric. Keep the 3:4 if coat length and face still match reality.

Using [YOUR PHOTO], show me in a [COAT TYPE] caught mid-stride in a studio cyc, 70mm at f/3.5, one hard side light so the fabric weave reads. Keep my face and the true length of the coat, including how it sits on my shoulders. I am alone. 3:4 fabric test, not a campaign hero.

💡

Pro tip: Say whether the coat is cropped or knee-length so hem does not invent itself.

Shoes, hem, and face

20/30

✨ What it does

ChatGPT frames your named shoes and hem in the lower third with your real face still readable above. Keep the 4:5 if the head was not cropped off.

From [YOUR PHOTO], make a picture that shows my [SHOE TYPE] and hem in the lower third, with my real face still readable in the upper third. 85mm at f/4, [FLOOR TYPE] floor, even overhead. Keep identity; do not crop my head off or swap in model legs. I stand alone. 4:5 outfit record.

💡

Pro tip: Name the floor (wood, concrete) so the shoes sit in a real space.

Yearbook looks

5 prompts

Classic yearbook bust

21/30

✨ What it does

ChatGPT restyles you as an adult yearbook bust on a mottled school backdrop. Keep the 4:5 if it is clearly you now, not a teenage recreation.

From [YOUR PHOTO], make an adult yearbook-style bust of me on a [BACKDROP COLOR] mottled school backdrop. 105mm at f/5.6, butterfly light, 4:5. I wear [WARDROBE]. Keep my real adult face and age; this is a playful restyle of me now, not a teen recreation and not a minor. I am the only person. Soft school-portrait vignetting.

💡

Pro tip: Say 'adult, my current age' so the model does not de-age the bust.

Club photo with a prop

22/30

✨ What it does

ChatGPT shows you in a yearbook-club style still holding a named prop in a simple room. Keep the 4:5 if your adult face stayed and no extra club members appeared.

Using [YOUR PHOTO], make a yearbook-club style picture of me holding [PROP], standing in a simple [ROOM TYPE] with a plain bulletin board behind. 50mm at f/4, fluorescent mixed with window. Keep my adult face; the prop is a hobby, not a costume identity swap. No other club members. 4:5.

💡

Pro tip: Pick a real hobby prop you own (camera, violin) so the grip looks natural.

Black sweater simple

23/30

✨ What it does

ChatGPT makes a quiet picture of you in a named black knit on dark gray seamless. Keep the 4:5 if hair and face stayed exact, with a calm expression.

From [YOUR PHOTO], make a simple yearbook-adjacent picture of me in a black [KNIT TYPE] on a dark gray seamless. 85mm at f/4, soft wrap, 4:5 from the collarbone up. Keep my hair and face exactly, including the neckline I have. Quiet expression. I am the only subject. Not a dating close-up.

💡

Pro tip: A true black knit tests whether the model keeps your neckline from the source.

Oak tree outdoor yearbook

24/30

✨ What it does

ChatGPT stands you under a named tree on a school-lawn look in open shade. Keep the 3:4 if it still reads as you as an adult, with an empty lawn.

Generate a picture of me from [YOUR PHOTO] standing under an [TREE TYPE] on a school-lawn look, 85mm at f/2.8, open shade, grass in the lower third. I wear [WARDROBE]. Keep my adult identity; this is a grown-up outdoor yearbook gag, not a teen scene. Empty lawn, no other students. 3:4.

💡

Pro tip: Open shade under the tree avoids raccoon eyes from overhead sun.

Three-pose contact sheet

25/30

✨ What it does

ChatGPT builds a three-pose contact sheet of you looking left, to camera, and right. Keep the 16:9 if the same face holds across all three frames.

From [YOUR PHOTO], build a three-pose contact sheet of me on [BACKDROP COLOR] paper: looking left, looking to camera, looking right. Same [WARDROBE], same 85mm, f/5.6. Keep my face consistent across the three frames. Adult only. No fourth person in any frame. 16:9 sheet.

💡

Pro tip: Ask for equal-sized frames so one pose does not dominate the sheet.

Creative studio

5 prompts

Paper roll color field

26/30

✨ What it does

ChatGPT sits you on a named paper roll with even high-key light. Keep the 4:5 if the color stays on the paper and your skin still looks like you.

Using [YOUR PHOTO], sit me on the sweep of a [PAPER COLOR] paper roll, knees tucked, 35mm at f/4, even high-key with a large overhead. I wear [WARDROBE]. Keep my face; the color field is the set, not a filter on my skin. I am alone on the paper. 4:5 studio still.

💡

Pro tip: If skin picks up too much paper color, ask for a white bounce near the face.

Projected pattern on a wall

27/30

✨ What it does

ChatGPT stands you in front of a named projected pattern with low ambient light. Keep the 3:4 if your face stays readable and the pattern does not become a mask.

From [YOUR PHOTO], stand me in front of a wall with a [PATTERN TYPE] projection, 50mm at f/2.8, low ambient, pattern stronger on the wall than on skin. I wear [WARDROBE]. Keep my face readable through the pattern; do not stencil the pattern onto my identity. I stand alone. 3:4.

💡

Pro tip: Keep the pattern on the wall more than on the face so identity holds.

Medium format film mock

28/30

✨ What it does

ChatGPT renders you as a square medium-format studio still with a named film look. Keep the 1:1 if face and age still match, with light grain only.

Generate a picture of me from [YOUR PHOTO] as if shot on a medium-format camera, [FILM LOOK] color, square 1:1, 80mm equivalent at f/4. I wear [WARDROBE] in a simple studio. Keep my real face and age. Light grain. I am the only person. Not a celebrity poster.

💡

Pro tip: Name a film look in plain words (warm portraits, cool daylight) not a brand slogan.

Cube sit, hard edge light

29/30

✨ What it does

ChatGPT sits you on a gray studio cube with hard edge light and a fill card. Keep the 4:5 if proportions stayed true and the cyc stayed empty.

From [YOUR PHOTO], sit me on a gray studio cube, hard edge light from the left, fill card on the right so the shadow side of the face still reads. 85mm at f/5.6. I wear [WARDROBE]. Keep my proportions; the cube is a set, not a throne. Empty cyc. 4:5.

💡

Pro tip: A fill card note stops the hard light from hiding half your face.

True profile, hard key

30/30

✨ What it does

ChatGPT makes a true profile of you with a hard key on dark paper. Keep the 4:5 if nose, chin and ear still match the source, not a statue.

Using [YOUR PHOTO], make a true profile picture of me, nose to the [PROFILE SIDE], hard key skimming the face, dark [BACKDROP COLOR] paper. 105mm at f/4, 4:5. Keep my real nose, chin and ear from the source; do not Greek-statue the profile. I am the only subject.

💡

Pro tip: If you only have a front source, say 'infer profile from this face, keep the same nose'.

Free tool

Prompt Optimizer

Turn a rough idea into a structured, professional AI prompt.

Try it free →

Frequently Asked Questions

Yes. These jobs start from [YOUR PHOTO]. A clear, front-facing picture with even light locks identity. Heavy filters and group crops make the face drift.
No. The set is for headshots, casual records, travel, outfit tests, yearbook gags and studio stills. Skip flirt copy and nightclub poses if you want the same tone.
Restate age, hairline, glasses and skin in the prompt. If a take looks like a stock model, rerun and say keep my real face, no beauty slim.
Yes. The outfit-test jobs swap [WARDROBE] while locking your face. Describe fabric and length. Check that the garment did not change your body.
Use 4:5 for most portraits, 3:2 for travel, 1:1 for film-style stills, 16:9 for splits and contact sheets. Match the ratio to where you will show the picture.