
Number the photos and name the roles
Write "image 1 is the scene, image 2 is the person to add". The prompt is the only place roles live, so be explicit about which is which.
Put two people together in a photo: upload the scene and a picture of the person to add, say where they should stand, and get one image with the light and scale matched.
Add a person to a photo with AI, or combine two people in one photo. Upload both pictures, say where they stand — light and scale matched. First run free.
Upload at least one reference image before generating.
Updated
All four take two photos and merge them. They differ in how well they match the light and the grounding shadow, so test a tricky scene on two.

OpenAI latest image model. Sharper instruction following, cleaner in-image text, up to 4K.

Photorealistic detail with tight prompt control. Accepts up to 8 reference images.

Highest quality Google model. Supports up to 4K resolution.

Strong typography and poster layouts. Handles dense in-image text at 1K and 2K.
A missing relative, a friend who could not come, a pet: the trick is making the new arrival look like they were in the room. Already have one shared scene photo? Insert the second person here. Starting from two separate portraits with no shared background? Use the couple photo maker instead.

Write "image 1 is the scene, image 2 is the person to add". The prompt is the only place roles live, so be explicit about which is which.

"At the right edge of the group, shoulder to shoulder with the woman in blue" is much better than "somewhere in the photo".

A person lit from the left dropped into a scene lit from the right is the giveaway. Add "match the light direction and the scale" to the prompt.

Someone with no shadow floats. Say "with a soft shadow on the ground in the same direction as the others".

A full-body shot from eye level adds into a similar scene far more convincingly than a selfie into a wide landscape photo.

Add "keep everyone in image 1 exactly as they are". Without it, faces nearby can shift while the model redraws the frame.

Use this for people who are happy to be in the picture. Do not use it to place someone in a scene they never agreed to, and be clear it is edited if you share it.
Same generator, pointed at a neighbouring job.
The step-by-step guide: phone, layered editor, or AI, plus five checks that make the result look real.
Combine several separate portraits into one family picture.
The general version for people, products, and style from up to five photos.
No shared scene photo? Build a couple shot from two separate portraits instead.
The reverse: take someone out and rebuild the space.
The general reference-image workspace for edits of any kind.
Upgrade for larger monthly credits, faster generation, and commercial-ready output.
Try GPT-IMG with signup credits
For regular creators and light production use
Generate 50–125 images / month
For high-volume image generation workflows
Generate 120–300 images / month
Buy credits once and use them whenever you need. Purchased credits never expire.
Practical details for this generator page.
Upload the scene and a photo of the person, number them in the prompt, say where the person stands, and ask for matching light, scale, and shadow with everyone else unchanged.
A reliable pattern: "Add the person from image 2 into image 1, standing at the right edge of the group. Match the scene light and scale. Keep everyone in image 1 exactly as they are." Adjust the position clause to your shot.
Yes. Upload the base photo as image 1 and the second person as image 2, then say where they should stand. The model redraws one frame with both people, matching light and scale.
Use this page when you already have a scene photo and want to insert someone into it. For two separate portraits with no shared background, use the AI couple photo maker instead.
Often, when the two photos have a similar angle and light and the prompt names both. Check the shadow and the edges, and re-run if either looks pasted.
Yes. Upload a photo of it and describe where it goes. The same light, scale, and shadow rules apply to objects and animals.
Yes for people who are happy with it, such as a relative who could not attend. Do not use it to put someone in a scene they have not agreed to.
New accounts start with 30 signup credits. A 1K Standard edit on GPT Image 2.5 is 15 credits, so your first two person added to a photo runs are covered. Images made on free signup credits carry a watermark.
Start with GPT Image 2.5, the default. FLUX 2 Pro costs the same 15 credits for a 1K Standard run, so you can compare the two on one brief. Light matching is the test, so run the same pair of photos through two models and keep the better-grounded result.
Use the prompt panel above with the same account, credits, and history workflow.