PromptPhoto editing

AI prompt to keep the same person in a new setting

Copy the prompt below, describe the new setting, and paste it with your photo into an AI image tool that accepts a reference photo. The prompt locks the person’s face, hair and proportions, then makes the light, shadows and scale agree with the scene so the result looks like one photograph. No prompt can promise an identical face, so generate a few versions and keep the closest.

Fill in and copy the prompt
1

Fill in the fields

The text on the right changes as you type.

Nothing you enter leaves your browser.

2

Copy the prompt

Your values are marked in colour.

Prompt17 lines
Use the attached photo as the identity reference. Create a new photograph of the same person in this setting: a bright co-working space with large windows and plants.

Scene details:
- time of day: late morning
- light: soft daylight from the left
- framing: waist-up portrait, 85mm lens look
- mood: confident and relaxed

Keep the person's face, hair, skin tone, age and body proportions identical to the reference. Keep their clothing the same unless I say otherwise.

Make the image agree with itself:
- the light on the person comes from the same direction and has the same softness as the light in the scene
- shadows fall on the ground and surfaces the way they would in that place
- the person's size fits the room, and their feet or seat touch a real surface
- the background has natural depth of field for the lens look above

Do not add other people, text, logos or watermarks. Make it look like a real photograph, not an illustration.

How to use it

  1. 1Choose a reference photo with a clear, front-facing face and no sunglasses or heavy shadows.
  2. 2Describe the setting in one sentence and pick one light direction. Fewer choices usually give more consistent results. If lens terms are new to you, write the framing as "waist-up portrait, softly blurred background".
  3. 3Paste the prompt with the photo, and generate a few versions.
  4. 4Compare each version to the reference. Keep the one where the face is closest, then fix the scene.

A good result looks like

A believable photo of the same person in the new place, with shadows and light that agree with the scene, and no sign that two images were joined.

Stay safe

  • Use photos of yourself, or of people who have agreed to the edit.
  • Do not use an edit to claim a job, license or rank the person does not have. A blazer is fine. A doctor’s coat on a non-doctor is not.
  • If a site asks you to say when a photo is edited, say so.
  • Keep a photo that still looks like you. Do not use an edit to change your face, body or age.
  • Do not create a photo that suggests a person was somewhere or did something they did not, if it could mislead anyone.

If it goes wrong

5 known problems, each with a fix

  1. 01

    Problem: The face drifts across versions.

    Fix: Use the same reference photo each time and add "same person, do not change facial features".

  2. 02

    Problem: The light does not match the scene.

    Fix: Say exactly one light source, such as "window light from the left".

  3. 03

    Problem: The person looks too large or small for the room.

    Fix: Describe the furniture next to them and the framing.

  4. 04

    Problem: The scene looks generic.

    Fix: Add two or three concrete objects that belong in it.

  5. 05

    Problem: The clothing changes between versions.

    Fix: Describe it in a few words, such as "same navy shirt", and repeat "keep their clothing the same".

Has it been tested?

Not tested yetmodel: none yet · date: none yet

Not yet. This one was written for this page and I have not run it on a specific model yet. Results will vary by tool, so run it once, check the result, and use the fix table above as your checklist. When I test it, the model, the date and the result will be listed here.

Questions

Why one light source?

Real photos usually have one dominant light. Models handle one clear source better than a list of several, and the result looks more believable.

Does this work for product photos?

The same idea works. Replace the person with the product, and keep the product’s shape, label and color fixed before describing the scene.

Why does the face still change when I say to keep it?

Many tools build a new image instead of moving your photo into a scene, so small changes to the face can appear. Use one clear reference photo, change as little else as you can, and keep the version that is closest.

Can ChatGPT, Claude or Gemini put me in a different place?

Some AI chat and image tools accept a reference photo and a description, and some do not. What each one can do changes often, so check the tool’s own help page. I have not tested this prompt on any model.

Build this for you

Want this built into your own workflow?

Tell me what you do by hand today. I will say what is safe to automate, what needs a person, and roughly what it would take. I reply within 24 hours.

Wondering what your project will cost?

Get a free price range and timeline in 24 to 48 hours. No commitment.

Get a Free Estimate