Skip to content

Couple Hug Photo Prompt (Demo)

Original photo placeholder (DEMO) for the couple hug prompt: abstract shapes, no real person
Original photo real, not AI
DEMO placeholder in the ChatGPT result slot of the couple hug prompt: abstract shapes, not AI output
AI-generated ChatGPT result

DEMO placeholder, not tested yet. Turns a photo of a couple standing side by side into a warm back hug in a mustard field at golden hour, both faces close together and turned to the camera. You need one clear photo of the two of you, both faces visible and about the same size.

  • Photos needed 1 couple photo
  • Model Best on ChatGPT
  • Face match Gemini Medium · ChatGPT High

Couple Hug Photo Prompt (Demo) prompt for Gemini and ChatGPT

ChatGPT prompt

KEEP: Edit the uploaded photo of two people; do not redraw their faces. Keep both faces, features, skin tones, hairstyles and ages exactly as they are in the photo. SCENE: A mustard field in full yellow bloom at golden hour. Pose, step by step: the person on the [LEFT/RIGHT] of the photo moves behind the other; their two arms wrap around the other's waist; their chin rests near the other's shoulder; both faces stay side by side and look at the camera. OUTFIT: The person in front wears a [KURTA COLOUR] cotton kurta with a phulkari-embroidered dupatta; the person behind wears a white kurta with a light brown shawl. LIGHT: Warm low sun behind them, both faces softly lit from the front; keep both skin tones exactly as in the photo. FRAME: Portrait 2:3, waist up, both faces fully visible and close together. AVOID: No text, no extra people, don't change face shape, don't lighten skin, don't swap or blend the two faces, no extra arms or hands, nothing suggestive.

DEMO: no real settings recorded (not a real test)

Tip: A photo of the two of you standing side by side, both faces clear

Our results: original vs Gemini vs ChatGPT

Result images are AI-generated with the model named in each caption and carry a burned-in "AI-generated · promptrang.com" label. The original is a real photo of our team, used with written consent. Outputs are shown as generated, only resized.

Original photo (real, not AI) · shared with consent
AI-generated with Gemini (DEMO placeholder) · tested 5 Nov 2026 · attempt 1 of 1 · no edits
AI-generated with ChatGPT (DEMO placeholder) · tested 5 Nov 2026 · attempt 1 of 1 · no edits
  • Gemini: Partial · face match Medium
  • ChatGPT: Pass · face match High

Verdict: DEMO: not a real test. On the real page this line gives the verdict from the test log, for example which model kept both faces closer and which one needed a fix line.

What photo to upload

Do

  • A photo of the two of you standing side by side, both faces clear
  • Both faces at about the same height and size in the frame
  • Even daylight on both faces
  • At least 1080 px on the short side

Don't

  • A photo that already shows a hug with one face hidden
  • Sunglasses, caps or beauty filters
  • A third person anywhere in the frame

Variations to try

Rain hug under a street light

Pose and setting: a side-on hug in a quiet lane on a rainy night.

KEEP: Use the uploaded photo of two people. Keep both faces, features, skin tones, hairstyles and ages exactly the same. SCENE: A quiet lane on a rainy night under a warm street light, light rain falling, wet cobblestones shining. The two hug side-on, cheek close to cheek, both faces turned towards the camera, smiling with eyes open. OUTFIT: The person on the left wears a [KURTA COLOUR] cotton kurta with a light jacket; the person on the right wears a dark shirt and a denim jacket. LIGHT: Warm street-light glow on both faces, cool blue rain around them; true-to-life colours and natural skin texture. FRAME: Portrait 4:5, chest up, both faces fully visible. AVOID: No text, no extra people, don't change face shape, don't lighten skin, don't swap or blend the two faces, no extra arms or hands.

Tested on Gemini.

AI-generated with Gemini · variation output · no edits

Railway platform reunion

Setting: an old railway platform in soft morning light, a happy reunion.

KEEP: Use the uploaded photo of two people. Keep both faces, features, skin tones, hairstyles and ages exactly the same. SCENE: An old railway station platform in soft morning light, iron pillars, a vintage maroon train carriage behind them, a suitcase at their feet. The two hug side-on in a happy reunion, both faces turned towards the camera. OUTFIT: The person on the left wears a [SAREE COLOUR] cotton saree; the person on the right wears a cream kurta with a brown sleeveless jacket. LIGHT: Soft morning light through the platform roof, gentle haze; true-to-life colours and natural skin texture. FRAME: Portrait 4:5, waist up, both faces fully visible. AVOID: No text, no extra people, don't change face shape, don't lighten skin, don't swap or blend the two faces, no signboards, station names or logos.

Tested on ChatGPT.

AI-generated with ChatGPT · variation output · no edits

Common problems and fixes

Extra arms or hands

Likely cause: DEMO: likely cause, not tested: Re-posing two bodies.

Fix: add this to your prompt.

The person behind uses exactly two arms around the waist; the person in front's hands rest on those arms.

One face turned away or hidden

Likely cause: DEMO: likely cause, not tested: Hugs hide faces.

Fix: add this to your prompt.

Both faces side by side, cheeks close, looking straight at the camera.

Skin looks lighter in the golden light

Likely cause: DEMO: likely cause, not tested: Glow effects.

Fix: add this to your prompt.

Keep both skin tones exactly as in the photo.

Test log

DateModel and versionPlanInputAttemptsResultTesterNotes
Gemini (DEMO placeholder)DEMO: no real photo used1PartialDEMO: no testerDEMO: not a real test
ChatGPT (DEMO placeholder)DEMO: no real photo used1PassDEMO: no testerDEMO: not a real test

FAQs

Can the AI make us hug if we're only standing side by side?

DEMO: placeholder answer. The real answer will come from the test log.

Who should stand behind?

DEMO: set [LEFT/RIGHT] in the prompt. Which choice gives cleaner results will come from the test log.