Candid smartphone photo prompts: 5 realistic phone shots on Z-Image and GPT Image 2.5

Z-Image Turbo prompt on Twistavo, an AI video studio for character swap: replace a person in a real video from 1 to 4 photos; plans from $3.99/month.

Updated

A photo that looks taken on a phone in an ordinary place reads as real: fluorescent light, a cluttered background, people who did not pose. We wrote five such prompts in September 2026, each with one absurd detail in a normal setting, and ran three on Z-Image Turbo and two on GPT Image 2.5. Here are the exact prompts, the settings, the credit cost, and the one where the model did the opposite of what we asked.

Make your own with Twistavo — 1 free image

Setup

Model
Z-Image Turbo
Settings
One vertical 9:16 image per prompt, text to image, no reference image. The words that make the photo look like a phone snapshot are the same in all five: "candid" or "casual phone photo", an ordinary place (laundrette, checkout, pavement), the real light of that place (fluorescent, overcast), real skin and cloth texture, and bystanders who react. Each prompt also says "original fictional" for every person. In Twistavo: open the AI Studio, pick Z-Image Turbo or GPT Image 2.5, paste the prompt, set 9:16 and the resolution shown, and generate. Z-Image Turbo costs a fraction of GPT Image 2.5 high (compare the cost under each example), so it suits trying many variations; GPT Image 2.5 follows long character descriptions closely.
Models used
GPT Image 2

1. Laundrette: a bodyguard presents one coin (Z-Image Turbo)

Realistic AI phone photo made with Z-Image Turbo: in a laundrette, a tall bald man in a black suit with an earpiece bows and holds a burgundy velvet cushion with one coin toward a tiny elderly woman in a lime-green coat and cream glasses, a wicker basket at her feet
  • Z-Image Turbo, text to image
  • Aspect ratio 9:16
  • Resolution 2k (1152x2048)

2 credits in Twistavo

Prompt

Candid realistic smartphone photograph inside a modest self-service laundrette. Two original fictional adults: a tiny elegant elderly woman with a sculpted silver beehive hairstyle, angular cream spectacles and a lime-green coat, calmly points to a coin-operated washing machine. Beside her an enormously tall broad-shouldered adult bodyguard in a plain black suit, clean-shaven bald head and discreet earpiece, bows gravely while presenting ONE small laundry coin on a burgundy velvet cushion. A wicker laundry basket is at her feet. The serious palace-level ceremony contrasts with the ordinary washing machines. Both full figures and faces visible. Real skin, natural fluorescent light, no fashion-ad gloss. Vertical photo, no lettering, brands, cartoon or watermark.

Tip: Every prop in the prompt is in the photo: the cushion, the single coin, the earpiece, the basket, the row of machines. Naming exact counts ("ONE small laundry coin") keeps the model from adding more. The fluorescent ceiling light does most of the realism.

2. Checkout: a weightlifter squat for one lemon (Z-Image Turbo)

Realistic AI phone photo made with Z-Image Turbo: in a supermarket aisle, a stocky man around sixty with a silver flat-top and a red velvet tracksuit squats like a weightlifter over a small paper bag holding one lemon while three shoppers watch
  • Z-Image Turbo, text to image
  • Aspect ratio 9:16
  • Resolution 2k (1152x2048)

2 credits in Twistavo

Prompt

Candid full-body smartphone photograph in an ordinary supermarket checkout area. An original fictional short stocky man around sixty, square clean-shaven face, very high silver flat-top hairstyle, huge serious eyebrows, red velvet tracksuit with narrow cream piping and white trainers. He has an intensely focused champion expression and a dramatic weightlifter squat, both hands reaching toward the handles of a comically tiny paper shopping bag containing ONE lemon on the floor. Nearby adult shoppers look quietly puzzled. His build is human and believable. Natural fluorescent supermarket light, real pores and cloth texture, documentary photography, understated absurd comedy. Vertical framing. No readable text, brand, watermark, illustration or cartoon.

Tip: The bystanders sell the joke: the prompt asks for "nearby adult shoppers look quietly puzzled" and the three people behind him do exactly that. The bag carries a few tiny printed marks; at phone size they read as noise, but check any surface where a brand could appear.

3. Wet sidewalk: a butler, an umbrella and a pigeon (Z-Image Turbo)

Realistic AI phone photo made with Z-Image Turbo: on a wet city sidewalk, a thin butler in a dark teal tailcoat and white gloves holds a black umbrella over his own head while a grey pigeon eats crumbs on the pavement beside him
  • Z-Image Turbo, text to image
  • Aspect ratio 9:16
  • Resolution 2k (1152x2048)

2 credits in Twistavo

Prompt

Candid full-body smartphone photograph on an ordinary wet city sidewalk. An original fictional tall thin middle-aged butler, long angular clean-shaven face, receding slick black hair, arched eyebrows, immaculate dark teal tailcoat, white gloves and oxblood shoes. With absolute seriousness he holds an elegant black umbrella very low over ONE normal grey pigeon eating crumbs from a small silver plate on the pavement. His own head is outside the umbrella getting wet. Both the man and the tiny bird clearly visible. Natural overcast light, damp wool, real skin pores, realistic street photography. Dry visual comedy, not glamorous advertising. Vertical composition. No writing, logo, watermark, cartoon or illustration.

Tip: This is the failure of the set. The prompt asked for the umbrella held low over the pigeon with the butler's own head in the rain; the model did the usual picture instead: the umbrella over the man, the pigeon in the rain. When the joke is a reversal, describe the positions of the objects in the frame ("the umbrella canopy is at knee height, above the bird") rather than the intention.

4. Supermarket aisle: a recurring character portrait (GPT Image 2.5)

Realistic AI phone photo made with GPT Image 2.5: waist-up portrait in a supermarket aisle of a fictional woman in her sixties with a huge crescent-shaped black hairstyle with a silver stripe, red rectangular glasses, a mustard-yellow suit and a burgundy bow
  • GPT Image 2.5 (sunburst), text to image
  • Aspect ratio 9:16
  • Quality high
  • Resolution 2k

20 credits in Twistavo

Prompt

Create ONE original fictional recurring social-media character in a mundane supermarket aisle, waist-up candid vertical phone photograph, face facing camera with deadpan confidence. Character: Colette, an elegant stocky French woman aged 62, broad square face, light olive skin, deep-set brown eyes, sharp arched eyebrows, long slightly hooked nose and thin unsmiling lips. Her real jet-black hair is arranged in a huge immaculate horizontal crescent-shaped coiffure sweeping outward above both ears like a black crescent moon, with a single narrow silver stripe centered at the hairline. Small rectangular red reading glasses, mustard-yellow sharply tailored skirt suit with very wide square shoulders, cream turtleneck, a rigid oversized burgundy silk neck bow. No props in hands, no gestures or visual gags. Her expression says that this astonishing appearance is completely normal to her. Strong readable human character silhouette, odd but endearing, eccentric not frightening. Real pores, fine age lines, authentic fabric, normal fluorescent supermarket lighting and shelves softly out of focus. Casual unretouched phone capture rather than cinematic or beauty advertising. One image only, no collage, no text, no watermark. Entirely original, no blond bob.

Tip: For a character you will reuse, GPT Image 2.5 followed the long description: the crescent hairstyle, the single silver stripe, the red glasses, the mustard suit. The shelves behind her are full of packaging; at web size nothing is readable, but at full 2k resolution look for real brand names before you publish.

5. Laundromat: a baguette held like a scepter (GPT Image 2.5)

Realistic AI phone photo made with GPT Image 2.5: in a busy laundromat, a fictional woman in her fifties with a silver-grey cloud-shaped bouffant and a scarlet 1970s suit sits upright holding a baguette vertically while two customers glance at her
  • GPT Image 2.5 (sunburst), text to image
  • Aspect ratio 9:16
  • Quality high
  • Resolution 2k

20 credits in Twistavo

Prompt

Photoreal vertical smartphone snapshot at a crowded self-service laundromat in Paris. Center frame: an original fictional French woman in her early 50s, olive skin, an enormous symmetrical silver-gray cloud-shaped bouffant with a sharp straight fringe, a completely deadpan expression, wearing a vivid scarlet 1970s tailored suit with exaggerated pointed white lapels and a single oversized brooch. She sits perfectly upright on a cheap plastic chair and holds one baguette vertically like a ceremonial scepter. Behind her, washing machines spin bright red socks; two ordinary customers in everyday clothes glance over, trying not to stare. The contrast between her theatrical, unmistakable silhouette and the mundane laundromat is immediately funny and slightly uncanny, but still plausible as a candid social-media moment. Straight-on phone-camera framing, a touch of harsh built-in flash, imperfect casual framing, rich natural skin and fabric texture, saturated but realistic colors, crisp readable subject, no editorial glamour and no movie-poster look. No lettering or watermark.

Tip: The same scene asks for the ordinary details that make a phone photo believable: red socks spinning in the machines, two customers trying not to stare, a cheap plastic chair. We then used this image as the base of a second prompt (an edit with an impossible reflection in the washer); a strong first image makes the edit easier.

Make it in Twistavo

  1. Open the AI Studio in Twistavo and pick Z-Image Turbo to try many ideas cheaply, or GPT Image 2.5 for a long character description.
  2. Paste a prompt and keep the phone-snapshot words: "candid" or "casual phone photo", an ordinary place, its real light, real texture, bystanders who react, and "original fictional" for every person.
  3. Change only the absurd detail, and write the positions of objects in the frame when the joke is a reversal.
  4. Set 9:16 and the resolution shown (2k; quality high for GPT Image 2.5), check the credit quote and generate. No reference image is needed.
  5. At full resolution, check shelves, bags and signs for readable brand names before you publish.

Variations to try

  • "Flash on, slightly overexposed" for a night snapshot.
  • A lower camera height: "photo taken from hip height".
  • The same character in three ordinary places, for a recurring series, as with the supermarket character on GPT Image 2.5.
  • The pigeon prompt rewritten with positions ("the umbrella canopy is at knee height, above the bird"): Z-Image Turbo put the umbrella over the man when the prompt only gave the intention.

Notes

Variations to try: the same prompt with "flash on, slightly overexposed" for a night snapshot, a lower camera height ("photo taken from hip height"), or the same character in three ordinary places for a recurring series. Every person on this page is fictional and was generated with AI with the prompt and settings shown. Results change from run to run.

Make your own with Twistavo — 1 free image

More AI video prompts

This prompt, its model, settings and result are part of the open dataset of every tested prompt (CSV and JSON).