The woman in this scenario opens Hinge at 10pm on a Sunday night with the bathroom ring light on, slides FaceTune to about 18% on her cheek and chin contour, and uploads four near-identical selfies from the same Tuesday in April. The match rate climbs. The date she meets on Friday spends the first ninety seconds quietly cataloguing the difference between the woman in his messages and the woman across the table. The seven prompts below are the fix, and they ship with one paste-line that stops the AI from making the same FaceTune-at-10pm mistake on her behalf.

What “filtered” actually costs on the date

The trade-off is real. A heavy filter and a ring-light at the right angle can move the match rate by a measurable margin on the right-swipe side. The same photos quietly tank the reply rate after the match, and they tank the date-show rate on the night itself. The published photo-feedback platforms have been clear about what the swiper is doing with the second-look. Photofeeler’s research blog has tracked this gap across years of peer-rated photo tests: photos that score higher on the “trustworthy” and “authentic” axes consistently out-perform photos that score higher on “attractive” but lower on those two. Variety wins the swipe. Authenticity wins the reply.

Hinge’s published profile-photo guidance names photo variety, a true full-body shot, and lead-photo eye contact as three of the strongest features it surfaces. None of those signals are filter-shaped. All three of them are what a Tuesday phone shot looks like, not what a Sunday-night ring-light selfie looks like.

The mechanism is older than the apps. OkCupid’s 2010 data team found that the lead photo with eye contact and a real smile carried a measurable lift over the same person in sunglasses or in a group. The post is fifteen years old and the underlying read has not changed: the fast-scrolling swiper has under a second of attention, and what they are reading is whether the face in the photo is the same kind of face that would walk into the bar on Friday.

The seven archetypes below are the answer to that read. The FaceTune-to-real downgrade rule is the line that keeps each one from drifting into ring-light territory.

The 7 female-coded archetypes

Each archetype below is rendered from one identity-reference selfie. The bone structure, the eye shape, the nose, the lips, and the freckles on the cheekbones hold across all seven. The wardrobe, the light, and the setting are what swing the read. The verbatim downgrade rule sits at the top of every prompt’s rules block: no FaceTune-style smoothing of cheek / chin contour; preserve natural facial micro-asymmetry. Pick the archetype the profile is missing, and paste. (Readers looking for the gendered counterpart can run the men’s seven-prompt pack instead; the male-default failure modes are different and the archetype names there reflect that.)

Archetype 1: the unposed-with-friends

Dating profile photo archetype 1, the unposed-with-friends: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones, caught mid-laugh at the edge of a sunlit outdoor picnic with two soft-bokeh friends in the background, identity-locked 4:5 vertical dating-app format showing the social-proof archetype that reads as a real Saturday afternoon rather than a posed photo.

Archetype 1: the unposed-with-friends. Edge of the group, not the centerpiece. Caught mid-laugh, friends in soft bokeh behind.

The unposed-with-friends sits at the edge of the group, not in the centre. Two friends in soft bokeh behind her, head turned back at the camera mid-laugh, the body half-in and half-out of the social frame. The signal it sends is the one a swiper cannot get from a solo close-up: she has actual friends, on actual Saturdays, in actual sunlight. Slot it second or third in the pack. Never lead with it. Lead photos with multiple faces tank match rate on every platform that has published the data.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: one clear, well-lit front-facing photo of your face.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image at the edge of a small group of {FRIENDS_COUNT} friends at {GROUP_SETTING}. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. The subject is at the right edge of the frame, head turned three-quarter back toward the camera, caught mid-laugh at something one friend said, eyes lit. The friends are in soft bokeh further into the frame, recognizable as people but not in focus. The subject wears {CASUAL_TOP}, hair slightly windblown with one flyaway at the temple. The setting shows real lived-in details, busy enough to read as real, never noisy enough to compete with the face. Soft warm late-afternoon natural sunlight from front-and-right, no harsh shadow, no editorial key. The subject occupies 55-65% of the vertical frame, framed from mid-thigh up. Skin shows visible pores, fine micro-texture, natural micro-asymmetry, any natural skin markings preserved, hair flyaways at the temples. Shot on a 35mm-equivalent lens at approximately f/2.8. Single 4:5 vertical dating-profile photo, identity-locked to the uploaded reference, the unposed-with-friends archetype.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference photo’s bone structure, eyes, nose, lips, proportions, and skin tone exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved — they are the anti-AI-plastic signal; do not smooth them out
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • The subject is at the edge of the group, not the centerpiece — she is caught mid-laugh, head turned back at the camera, not posed as the focal point
  • Friends visible in soft bokeh further into the frame; they must NOT be in focus; their faces should be partially turned away or in profile
  • Soft warm natural late-afternoon sunlight only: no harsh shadow, no editorial cinematic key, no Vogue stylization, no ring-light flatness
  • No text, captions, watermarks, logos, brand marks, dating-app UI chrome, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: one clear, well-lit front-facing photo of your face
  • {FRIENDS_COUNT} = two, OR one, OR three
  • {GROUP_SETTING} = a sunlit outdoor picnic blanket in a city park on a Saturday afternoon, OR a sunlit back patio on a weekend, OR an outdoor cafe two-top on a quiet afternoon
  • {CASUAL_TOP} = a soft chambray shirt unbuttoned over a plain white tee, OR a faded olive-green knit sweater, OR a heather-grey relaxed-fit tee

Archetype 2: the at-home-not-glam

Dating profile photo archetype 2, the at-home-not-glam: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones sitting cross-legged on a linen couch in an oatmeal cardigan with a stack of books on her lap, soft Sunday-morning window light, identity-locked 4:5 vertical dating-app format showing a real lived-in living room rather than a glam-room mirror.

Archetype 2: the at-home-not-glam. Real couch, real cardigan, real books. Not the vanity, not the bathroom mirror.

The at-home shot is where most female profiles go wrong on the second tile. The default move is the glam-room mirror, the curated vanity, the bathroom selfie with the ring light propped on the sink. The at-home-not-glam goes the other way: a worn linen couch, an oatmeal knit cardigan, a real stack of books on the lap, the corner of a real fiddle-leaf-fig in soft focus. The swiper reads “this is what her Sunday morning actually looks like” instead of “this is what her bathroom mirror looks like.” That distinction is the entire move.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image sitting cross-legged on a worn linen couch in their own living room on a quiet Sunday afternoon. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They are in a three-quarter pose, body angled toward camera at about twenty degrees, looking slightly off-camera toward an unseen window, caught mid-soft-smile, not posed. They wear {HOME_LAYER}, with sleeves pushed to mid-forearm, hair in a half-up clip with two flyaways visible at the temples. A real stack of three books rests on their lap; they hold the top one half-open with one hand. The background shows {ROOM_DETAILS}; the room is busy enough to read as their actual home, never a styled rental or showroom. Soft natural late-morning window light from the right side, warm in tone, with gentle fall-off into the left cheek shadow; no studio key, no editorial directional light. The subject is NOT in a bathroom mirror, NOT in a glam-room mirror, NOT facing a vanity; this is a real lived-in living room. The subject occupies 60-70% of the vertical frame. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways at temples, natural micro-asymmetry. Shot on a 50mm-equivalent lens at approximately f/2.8. Single 4:5 vertical at-home-not-glam dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The setting must read as the subject’s real lived-in home, not a styled rental, not a glam vanity, not a bathroom mirror, not a showroom: real plants, real stack of books, lived-in surfaces
  • Soft natural window light only: no studio key, no editorial directional light, no ring-light flatness
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only: solo subject
  • No text, captions, watermarks, logos, brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {HOME_LAYER} = a soft oatmeal knit cardigan over a plain heather-grey tee, OR a worn olive sweatshirt over a white tee, OR a soft beige linen button-down over a tank
  • {ROOM_DETAILS} = the corner of a real living room with one fiddle-leaf-fig plant slightly out of focus, a soft-bokeh stack of magazines on a side table, a single warm-amber framed print on the wall; OR a soft-bokeh bookshelf with a small ceramic vase and a thrifted lamp; OR a corner of the room with a worn rug, a side table with a tea mug, and the edge of a soft throw blanket

Archetype 3: the morning-coffee-window

Dating profile photo archetype 3, the morning-coffee-window: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones standing at a sunlit kitchen window in a plain white tee, holding a beige ceramic mug of coffee in both hands, looking down into the mug with a quiet half-smile, soft warm late-morning window light, identity-locked 4:5 vertical dating-app format.

Archetype 3: the morning-coffee-window. Mid-thought, looking down into the mug. Not staring at the camera.

The morning-coffee-window is the close-cropped, eye-line-down variant that works as a lead photo on Hinge. The reason it lands as the lead is that the gaze is not staring into the camera; it is staring into the mug. Warm window light from the front-left, a quiet half-smile, hair tucked behind one ear. The swiper reads “caught mid-thought” instead of “posed for the camera.” It is the closest thing to a candid that a solo phone-render can produce, and it ships the eye-contact-and-smile signal that OkCupid’s 2010 data named as the largest single lift over a sunglasses-or-group lead.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image standing at a sunlit kitchen window in their own apartment on a quiet late morning, holding {MUG_DETAIL} in both hands at chest level. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They are in a three-quarter pose facing the window, body angled twenty degrees away from camera with the head turned softly back over the shoulder, looking down into the mug with a quiet half-smile (not staring into the camera, not posing — caught mid-thought). Hair is tucked behind one ear with two flyaways at the temple. They wear {COFFEE_OUTFIT}. The light source is the kitchen window directly in front: soft natural late-morning warm window light from front-and-left, with a hint of warm reflection on the left cheek and gentle fall-off into the right cheek shadow. The kitchen is real and lived-in: {KITCHEN_DETAILS}. The subject occupies 60-70% of the vertical frame. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways, natural micro-asymmetry. Shot on a 50mm-equivalent lens at approximately f/2.8. Single 4:5 vertical morning-coffee-window dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The setting must read as a real lived-in kitchen, not a styled showroom, not a coffee-shop bar; the window must be a real apartment window with real morning light
  • Soft natural warm late-morning window light only: no studio key, no editorial directional light, no ring-light flatness, no golden-hour stylization
  • The subject is caught looking down into the mug with a quiet half-smile, not staring into the camera; this is a mid-thought morning shot, not a posed portrait
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only: solo subject
  • No text, captions, watermarks, logos, brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {MUG_DETAIL} = a beige ceramic mug of coffee, OR a heavy stoneware mug of black tea, OR a small white espresso cup
  • {COFFEE_OUTFIT} = a plain white relaxed-fit tee tucked into soft camel wide-leg trousers, OR a heather-grey crew-neck sweater over dark jeans, OR a soft oversized button-down over pajama-style soft pants
  • {KITCHEN_DETAILS} = a soft-bokeh edge of a wooden counter with a small white ceramic spoon-rest, a real eucalyptus stem in a glass jar to the left, the corner of a real fridge with one small magnet visible further back; OR a soft-bokeh window sill with a small potted herb plant, a folded dishtowel over the counter edge

Archetype 4: the hobby-mid-laugh

Dating profile photo archetype 4, the hobby-mid-laugh: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones, mid-laugh while throwing pottery at a wheel in a sunlit ceramics studio, hands wet with grey clay, wearing a paint-flecked denim apron over a charcoal tee, identity-locked 4:5 vertical dating-app format showing the caught-mid-hobby archetype that gives a swiper a real conversation hook.

Archetype 4: the hobby-mid-laugh. Hands wet with clay, mid-laugh at the wheel. The conversation hook on a plate.

Hinge’s profile-photo guidance specifically calls out photos showing a hobby or interest as one of the lift drivers on first-message rates, not just on match rates. The mechanism is simple: a close-up smile says “I am a person.” A photo of her mid-laugh at the pottery wheel, hands wet with grey clay, says “I throw pottery on Tuesdays, here is your opening line.” The hobby itself is interchangeable. Swap the pottery for an easel, an acoustic guitar on the couch, a yoga mat mid-class, a sourdough proof on the kitchen counter. The rules block holds the rest constant.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image caught mid-laugh while {HOBBY_VERB}. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They are mid-action with {HOBBY_BODY_POSITION}, body slightly angled away from camera at about thirty degrees, head turned back toward camera in a real spontaneous laugh, eyes lit and crinkled at the corners. They wear {HOBBY_OUTFIT}, sleeves pushed to mid-forearm; hair is half pulled back with two flyaways at the temple. The backdrop is {HOBBY_SETTING}, real and lived-in, with slightly out-of-focus context cues visible in soft bokeh further back. The light is soft warm setting-appropriate natural light, no harsh shadow, no editorial cinematic key, no ring-light. The subject occupies 60-70% of the vertical frame. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways, a faint flush on the cheeks from concentration and laughter, natural micro-asymmetry. Shot on a 50mm-equivalent lens at approximately f/2.8. Single 4:5 vertical hobby-mid-laugh dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The image must read as caught mid-action, not posed: hands placed naturally on real props, body weight as someone actually doing the hobby, expression a real mid-laugh rather than a staged smile
  • Soft warm setting-appropriate indoor or outdoor natural light only: no editorial cinematic key, no Vogue stylization, no ring-light flatness
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, faint exertion flush on cheeks, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only as the sharp subject
  • No text, captions, watermarks, logos (including any brand logos on apron, tools, or equipment), brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {HOBBY_VERB} = throwing pottery at a wheel with hands wet in grey clay, OR painting at an easel mid-brushstroke, OR strumming a chord on an acoustic guitar on a couch, OR mid-pose at a yoga mat in a sunlit studio, OR shaping sourdough on a floured wooden counter
  • {HOBBY_BODY_POSITION} = both hands shaping the rising walls of a real clay cylinder with wet clay smudges on the wrists, OR one hand mid-brushstroke on the canvas with the other holding the palette, OR one hand on the fretboard with the other mid-strum
  • {HOBBY_OUTFIT} = a paint-flecked denim apron over a charcoal cotton tee, OR a soft heather-grey tee with rolled cuffs, OR a fitted black tank with soft loose pants for yoga
  • {HOBBY_SETTING} = a sunlit ceramics studio with an unfinished bisque pot on a wooden shelf, a row of metal sculpting tools, a small bag of clay; OR a sunlit home art corner with a real canvas and palette; OR a soft-bokeh real living room with a lived-in side table for guitar

Archetype 5: the walking-not-standing

Dating profile photo archetype 5, the walking-not-standing: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones, mid-stride walking along a tree-lined city sidewalk in a long camel coat over a charcoal turtleneck, dark jeans, and broken-in white sneakers, full body visible from mid-thigh up, soft overcast late-afternoon daylight, identity-locked 4:5 vertical dating-app format showing the walking-not-standing archetype that delivers a real full-body shot without a tripod or photographer.

Archetype 5: the walking-not-standing. Mid-stride, one heel lifting. The full-body shot that doesn't read as posed.

Hinge has named the full-body shot as one of the photo types most missing from underperforming profiles. The lift from adding a single full-body shot to a profile that does not yet have one is, by Hinge’s own published guidance, larger than the lift from any other single change. The default render of “full-body shot” is a posed-on-a-staircase composition that reads as a magazine spread. The walking-not-standing breaks that default: one heel lifting off the pavement, body weight forward, head turned just off-camera at something the viewer cannot see. Mid-stride is the rule. Soft overcast daylight with dappled sun through the trees is the lighting. A long camel coat over dark jeans is the wardrobe-default. The result is a full-body shot that reads as a walk on a Tuesday afternoon, not a shoot.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image walking along {WALKING_LOCATION} on a soft-overcast late afternoon, full body visible from mid-thigh up. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They are mid-step, with one heel just lifting off the ground, body weight forward, three-quarter angle to camera. The head is turned just off-camera to the right, caught mid-smile at something the viewer cannot see — a passing dog, a friend across the street, a window display. They are NOT standing still posed for the camera; they are walking, mid-stride. They wear {WALKING_OUTFIT}. {WALKING_ACCESSORY}. Hair slightly wind-lifted at the back, two flyaways at the temple. The sidewalk is a real city street: real autumn leaves scattered on the pavement, a soft-bokeh row of brownstone storefronts further back, dappled sunlight broken by the trees overhead. Soft overcast daylight from above with cool tone in the highlights, no harsh shadow, no editorial cinematic key, no golden-hour stylization. The subject occupies 60-70% of the vertical frame and is shown from mid-thigh up so the full-body shot signal is unmistakable. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways, slight pink in cheeks from the cool air, natural micro-asymmetry. Shot on a 35mm-equivalent lens at approximately f/4. Single 4:5 vertical walking-not-standing dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The subject MUST be mid-step walking, not standing still; one heel lifting, body weight forward, mid-stride
  • Full-body framing required: subject shown from mid-thigh up at minimum; this is the full-body archetype and must read as such at a glance
  • Soft overcast diffused daylight with dappled sunlight through trees only: no harsh sun, no editorial key, no golden-hour stylization, no ring-light flatness
  • The scene must read as a real city sidewalk they actually walk on: real terrain, real autumn leaves, real storefront context in soft bokeh, broken-in shoes
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, slight cool-air flush in cheeks, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only: solo subject
  • No text, captions, watermarks, logos, brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {WALKING_LOCATION} = a tree-lined city sidewalk on a regular weekday, OR a brownstone-lined neighborhood street, OR a quiet farmers-market street in the morning
  • {WALKING_OUTFIT} = a long camel wool coat over a charcoal turtleneck, dark slim-cut jeans, and broken-in white sneakers; OR a soft denim jacket over a heather-grey tee, dark jeans, and ankle boots; OR a soft mid-length trench over a cream knit, dark jeans, and white sneakers
  • {WALKING_ACCESSORY} = a worn canvas tote bag with a single olive-green strap hangs from the left shoulder, with the right hand in the coat pocket; OR a soft leather crossbody bag with one hand resting at the strap; OR sleeves of the coat pulled to the wrists with both hands free at the sides

Archetype 6: the kitchen-not-mirror

Dating profile photo archetype 6, the kitchen-not-mirror: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones, mid-laugh while chopping a red bell pepper at her own kitchen island on a quiet weeknight, in a heather-grey sweatshirt with sleeves pushed to mid-forearm, warm overhead pendant light, identity-locked 4:5 vertical dating-app format showing the candid kitchen archetype rather than a bathroom-mirror selfie.

Archetype 6: the kitchen-not-mirror. Real knife, real cutting board, real weeknight. Not the bathroom selfie.

The kitchen-not-mirror is the answer to the bathroom-mirror-selfie habit. The mirror-selfie default is the single most-flagged failure mode in the female dating-photo subforums. The fix is not “take a better mirror selfie.” The fix is “do not be in a mirror.” The kitchen-not-mirror puts her at the kitchen island chopping a red bell pepper on a wooden board, knife mid-slice, mid-laugh up at the camera, in a heather-grey sweatshirt with the sleeves pushed to mid-forearm. The warm overhead pendant light replaces the ring light. The bowl of chopped onions on the counter replaces the soap bottle in the mirror frame. Same face, opposite read.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image standing at their own kitchen island chopping {COOKING_INGREDIENT} on a wooden cutting board on a quiet weeknight. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They stand at the kitchen island with both hands on the cutting board: the right hand holds a real chef’s knife mid-slice, the left hand holds the half-chopped ingredient steady. They are mid-laugh, head turned slightly up and toward the camera, eyes lit and crinkled. They wear {COOKING_OUTFIT}, hair in a messy low ponytail with two flyaways at the temple. A real lived-in kitchen surrounds them: {COOKING_KITCHEN_DETAILS}. They are NOT in a bathroom mirror, NOT in a glam-room mirror, NOT at a vanity — this is a real working kitchen island. The light source is the warm overhead pendant above the island plus soft natural light from a window to the right: warm in tone, gentle fall-off into the left cheek shadow, no harsh shadow, no editorial cinematic key, no ring-light. The subject occupies 60-70% of the vertical frame. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways, natural micro-asymmetry, a slight evening flush on the cheeks. Shot on a 50mm-equivalent lens at approximately f/2.8. Single 4:5 vertical kitchen-not-mirror dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The subject is at a real working kitchen island chopping a real vegetable, NOT at a bathroom mirror, NOT at a glam-room vanity, NOT at a posed kitchen-counter selfie angle; this is a candid weeknight cooking moment
  • Warm overhead pendant light + soft natural window light only: no studio key, no editorial directional light, no ring-light flatness, no Forbes-cover gravitas
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, slight evening flush, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only: solo subject
  • No text, captions, watermarks, logos (including any brand text on cookbook cover), brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {COOKING_INGREDIENT} = a red bell pepper, OR a bunch of fresh herbs, OR a yellow onion, OR a wedge of butternut squash
  • {COOKING_OUTFIT} = a heather-grey crew-neck sweatshirt with sleeves pushed to mid-forearm, OR a soft olive-green long-sleeve tee with rolled cuffs, OR a faded denim button-down over a white tank with sleeves rolled
  • {COOKING_KITCHEN_DETAILS} = a real ceramic bowl with chopped onions and trimmings on the counter, a half-full clear glass of water, a wooden spoon resting on a folded dishtowel, a small open cookbook propped against the toaster in soft bokeh, the corner of a stainless fridge visible behind the shoulder; OR a small wooden cutting board with garlic skins, a half-poured glass of red wine on the counter, a small herb plant on the windowsill

Archetype 7: the outdoor-not-vacation

Dating profile photo archetype 7, the outdoor-not-vacation: a 30-year-old woman with shoulder-length brown hair, brown eyes, and slight freckles on her cheekbones sitting on the stone steps of a city library with a takeaway coffee and a paperback book on her lap, in a charcoal wool turtleneck and dark jeans, soft overcast weekday late-afternoon daylight, identity-locked 4:5 vertical dating-app format showing a real city-life full-body shot rather than a tropical-vacation cliché.

Archetype 7: the outdoor-not-vacation. City library steps, takeaway coffee, weekday afternoon. Not Bali.

The outdoor-not-vacation kills the Bali-beach-throwback-from-2019 habit. A vacation photo on a dating profile sends two signals at once, and only one of them is the one the woman is trying to send. The good signal is “I have a life.” The bad signal is “the best version of me is the one that was on holiday three years ago.” The outdoor-not-vacation pulls the outdoor shot into her own city on a regular weekday: stone library steps, a takeaway coffee, a paperback open on the lap, a charcoal wool turtleneck. Same outdoor signal. None of the Bali baggage.

Show the full promptTap to expand

Paste this into your AI (ChatGPT, Claude, Gemini, or any AI image tool).

REQUIRED upload before pasting: the same reference selfie of your face you used for archetype 1.

Generate this image:

A single photoreal 4:5 vertical dating-profile photo of the person in the uploaded reference image sitting on {CITY_LOCATION} on a quiet weekday afternoon, a takeaway coffee cup beside them and a paperback book open on the lap. Match their bone structure, eyes, nose, lips, proportions, and skin tone exactly so the subject is unmistakably the same person from any earlier photo in the pack. They are seated on {CITY_SEATING}, knees pulled in slightly toward the body, three-quarter angle to camera. The head is turned just off-camera to the right, caught mid-soft-laugh at something the viewer cannot see. The right hand rests on the open paperback, holding the page; the left hand holds the disposable coffee cup at the side. They wear {CITY_OUTFIT}. Hair slightly windblown at the back with two flyaways at the temple. The setting is real and city-specific: {CITY_DETAILS}. They are NOT in a vacation setting — no beach, no famous landmark, no resort, no tropical-destination cliché; this is their own city, on a regular weekday afternoon, in a place they actually go. Soft overcast late-afternoon daylight from above with a cool tone, no harsh shadow, no editorial cinematic key, no golden-hour vacation stylization. The subject occupies 60-70% of the vertical frame and is shown from mid-thigh up. Skin shows visible pores, fine micro-texture, any natural skin markings preserved, hair flyaways, slight cool-air flush in cheeks, natural micro-asymmetry. Shot on a 50mm-equivalent lens at approximately f/4. Single 4:5 vertical outdoor-not-vacation dating-profile photo, identity-locked.

Rules the AI must follow:

  • Aspect ratio 4:5 vertical: strict, locked at the start and the end of the prompt
  • Identity preservation is the highest-priority constraint: match the uploaded reference exactly
  • Any natural skin markings (freckles, moles, scars) visible in the reference MUST be preserved
  • No FaceTune-style smoothing of cheek or chin contour; preserve natural facial micro-asymmetry
  • The setting MUST read as the subject’s own city on a regular weekday — a real library step, a real city street; NOT a vacation setting, NOT a beach, NOT a famous landmark, NOT a resort, NOT a tropical-destination cliché
  • Soft overcast late-afternoon daylight only: no harsh sun, no editorial key, no golden-hour stylization, no ring-light flatness, no travel-magazine cover light
  • Full-body framing from mid-thigh up: the outdoor-not-vacation archetype is a weekday-life full-body shot
  • Realistic skin texture required: visible pores, fine micro-asymmetry, hair flyaways, slight cool-air flush, natural unevenness; no porcelain smoothing, no over-retouched beauty filter, no waxy AI-plastic surface
  • One human figure only as the sharp subject (soft-bokeh pedestrians in deep background are acceptable but must NOT be in focus)
  • No text, captions, watermarks, logos (including any brand text on coffee cup, book cover, or shoes), brand marks, badge text, or readable signage anywhere in the frame
  • Single image output: one 4:5 file; no contact sheet, no variant grid
  • Output the image directly without explaining the prompt back

Replace these placeholders with your details:

  • REQUIRED upload before pasting: the same clear front-facing selfie you used for archetype 1
  • {CITY_LOCATION} = the stone steps of a city library, OR a wooden bench in a small neighborhood park, OR the low stone wall outside a museum
  • {CITY_SEATING} = a worn limestone step, OR a weathered wooden bench, OR a low stone garden wall
  • {CITY_OUTFIT} = a charcoal wool turtleneck, dark indigo straight-leg jeans, and worn brown leather ankle boots; OR a long camel coat over a cream sweater with dark jeans; OR a soft olive jacket over a heather-grey tee with dark jeans
  • {CITY_DETAILS} = real city pigeons on the steps further down in soft bokeh, the edge of a stone column to the left, a soft-bokeh blur of city pedestrians walking past in the deep background, real fallen autumn leaves on the steps; OR a soft-bokeh row of brownstones across the street, a real city tree casting dappled light, a passing cyclist out of focus

One short detour before the downgrade rule. The seven archetypes above sit inside a broader pack of twenty-five visual jobs that come up repeatedly in a normal year: the LinkedIn headshot, the founder portrait, the Etsy listing photo, the personal-brand grid, and twenty-one others. The Independent Brand Visual Kit is the free starter pack we send new subscribers: twelve copy-ready AI image prompts plus one paste-ready AI move a week after that. The dating archetypes live in the full library; the starter kit covers the photo jobs most readers paste next.

The FaceTune-to-real downgrade rule

Every archetype above inherits one line in its rules block. The line is the FaceTune-to-real downgrade rule:

no FaceTune-style smoothing of cheek / chin contour; preserve natural facial micro-asymmetry.

That single sentence is the named primitive. It is what stops the AI model from defaulting to porcelain skin, mirror-symmetric face, and a sharpened jawline. The model is trained on more filtered selfies than unfiltered ones; the dominant signal in its training distribution is the FaceTune look, and the only reliable override is an explicit named-failure-mode constraint. “Make it natural” does not work. “No filter” does not work. The downgrade rule works because it names the specific thing the model defaults to and tells it not to do that.

The before-and-after pair below shows what the rule does and what happens without it. The two images are the same persona at the same identity-reference photo. The only difference is whether the downgrade rule was in the prompt.

The BEFORE side of the FaceTune-to-real downgrade comparison: a 30-year-old woman in a tight black knit top taking a heavily-filtered bathroom-mirror selfie at 10pm with a ring-light visibly in the frame, showing the failure mode this article defeats (porcelain-smoothed skin with no pores, mirror-symmetric face, sharpened jawline, freckles entirely smoothed away, glassy ring-light reflection in the pupils), the canonical heavy-FaceTune dating-profile lead photo.

Before: no downgrade rule. Porcelain skin, mirror-symmetric face, ring-light flat front-light, freckles smoothed away.

The AFTER side of the FaceTune-to-real downgrade comparison: the same 30-year-old woman from the BEFORE image but rendered with the downgrade rule applied, slight freckles on cheekbones visible and preserved, no cheek or chin contour smoothing, natural facial micro-asymmetry preserved, soft natural window light replacing the ring-light flatness, phone-held casual selfie in a real lived-in living-room corner rather than a bathroom mirror, identity-locked 4:5 vertical dating-app lead-photo format demonstrating what an AI render looks like when the FaceTune-to-real paste-line is in the rules block.

After: downgrade rule applied. Freckles preserved, natural micro-asymmetry, soft window light, real living-room corner.

The methodology behind why this works lives in why AI images look fake. The short version: the model has learned that “attractive” photos are filtered photos, because most of the training data is filtered. The downgrade rule overrides that learned default by naming the specific failure mode. The rule does the work. The reader just has to not paste a softened version of it.

Assembling the six-photo pack from the seven archetypes

The pack is six photos, not seven. The job is to pick six of the seven archetypes for the six profile slots, plus one real group photo from the camera roll. The table below maps each archetype to the slot it fits, the algorithm signal it ships, and whether it can lead.

SlotRecommended archetypeAlgorithm signal it fillsCan it lead?
1 (lead)The morning-coffee-window (or close-cropped hobby-mid-laugh)Lead-photo eye-line + warm half-smile + face occupies 55-65% of frameYes: designed for it
2The at-home-not-glamPhoto variety + lived-in interior signalNo: better as second-tile shift
3The walking-not-standingFull-body shot (Hinge’s largest single lift) + photo varietyNo: full-body is rarely the lead
4The hobby-mid-laughFirst-message hook + photo varietyYes: works as lead on hobby-led profiles
5The kitchen-not-mirror or the outdoor-not-vacationPhoto variety + candid-life signalNo
6One real group photo from the camera rollSocial proof, actually has friendsNo: group photos as lead tank match rate

The unposed-with-friends is the swap candidate. Run it instead of one of the slot-2-to-5 archetypes if her real group-photo-from-camera-roll for slot six is weak or unclear. Both the unposed-with-friends and the slot-6 real group photo can sit in the pack at once, but two friends-shots in a six-slot pack is the upper limit. Three or more friends-shots and the swiper starts asking which face is hers. The gender-neutral five-photo dating pack at dating profile photo examples covers the slot logic at the cross-gender level if a different reader needs the unfiltered baseline.

The one decision the reader has to make is which photo from her camera roll becomes the identity-lock reference upload. That single photo is what every archetype inherits. The freckles on the cheekbones in every render above are inherited from the reference photo, not invented by the model. Pick the version of her face that she walks around as today, even if the lighting is mediocre. The AI fixes the lighting. The wrong reference identity, anchored into seven renders, cannot be fixed after the fact.

FAQ

Q: Will the same seven-archetype pack work on Hinge, Bumble, and Tinder for women?

A: Yes. The seven archetypes (the unposed-with-friends, the at-home-not-glam, the morning-coffee-window, the hobby-mid-laugh, the walking-not-standing, the kitchen-not-mirror, the outdoor-not-vacation) cover every signal Hinge, Bumble, and Tinder have publicly named as a lift driver for women: photo variety, full-body, a real hobby, lead-photo eye contact. The slot order is the only thing that changes. Lead with the morning-coffee-window or a close-cropped variant of the hobby-mid-laugh on Hinge (lead-photo weight is highest there). The unposed-with-friends is slot two or three, never lead, because lead photos with multiple faces tank match rate on every platform that has published the data.

Q: Why is the FaceTune-to-real downgrade rule the named primitive, not “use less filter”?

A: “Use less filter” is advice, not a rule. The downgrade rule is the literal line you paste into the rules block of the prompt so the AI model stops the smoothing at generation time, not at edit time. The verbatim line is “no FaceTune-style smoothing of cheek / chin contour; preserve natural facial micro-asymmetry.” Pasting that one line is the difference between a render that looks like a Tuesday phone shot and a render that looks like a ring-light selfie at 10pm. It works because the model honors named-failure-mode constraints more reliably than it honors vague aesthetic directives.

Q: What if my freckles or other natural features get smoothed away by the AI by default?

A: That is the failure mode the downgrade rule defeats. Name the specific feature you want preserved in the prompt: “slight freckles on cheekbones, preserve; do not smooth.” The model defaults to a 22-year-old porcelain skin tone because that is the dominant signal in its training data; the only thing that overrides it is the explicit named instruction. Identity-lock against your real reference photo carries the underlying bone structure across; the downgrade rule carries the texture.

Q: Is AI-generated dating profile photos allowed by the apps?

A: The community guidelines for Hinge, Bumble, and Tinder forbid catfishing, which is photos of someone who is not you or photos heavily edited to misrepresent how you look. Identity-locked AI renders of your own face, in clothes you actually wear, in scenes that match the kind of life you actually live, do not cross that line. The line is misrepresentation, not generation method. The photo that breaks the rule is the one where the date does not recognize you when you walk in. The seven archetypes here are written to defend that read.

Key Takeaways

  • Variety wins the swipe; authenticity wins the reply. Photofeeler’s research has tracked the “trustworthy and authentic” axes out-performing “attractive” in years of blind testing.
  • Seven archetypes, one face: the unposed-with-friends, the at-home-not-glam, the morning-coffee-window, the hobby-mid-laugh, the walking-not-standing, the kitchen-not-mirror, the outdoor-not-vacation. Pick six.
  • The FaceTune-to-real downgrade rule is one paste-line: “no FaceTune-style smoothing of cheek / chin contour; preserve natural facial micro-asymmetry.” It goes in every archetype’s rules block.
  • The pack ships in one evening. One identity-reference selfie. Six rendered slots. One real group photo from the camera roll. The date-day reveal is the only quality bar that matters.

The next Sunday night

The woman in the opening scene opens her app again the following Sunday. The forty-minute swipe-batch still climbs. The difference is on Friday night: the date who walks in stops counting the seconds before the recognition lands, and the conversation starts on the pottery wheel and the camel coat instead of on the silence that follows the seat-pulled-out moment. The algorithm has not changed. The pack has.

Which archetype is missing from her six? The honest answer is whichever one is not already on the profile. Pick that prompt first, render it tonight, drop it into the slot tomorrow. The seven dating archetypes here are the small subset of the $19 Image Prompt Pack (twenty-five paste-ready prompts: dating, LinkedIn, founder portrait, Etsy listing, personal-brand grid, and twenty more) that women specifically paste first.