How to add a person to a photo, three ways
Somebody was behind the camera, arrived late, or could not be there at all. There are three honest ways to add a person to a photo — an AI generator, a phone editing app, or a desktop editor with layers — and they are good at genuinely different things. This page says which to pick and shows the one we run ourselves.
Three ways to add a person to a photo
They are not ranked. They fail differently, which is what should decide it.
1 — AI generator (best when the light does not match)
Upload the group shot and a portrait of the missing person, describe where they stand, and the model redraws the whole frame with everyone lit the same way. Strongest method when your two sources were shot in different light, because it relights rather than pastes. Weakest when you need the original photo pixel-for-pixel untouched.
2 — Phone app (best when you are in a hurry)
Cut-out tools in most phone editors will lift a person and drop them into another shot in about a minute. Fine for something casual. It moves pixels without relighting them, so the join gives itself away the moment the two photos disagree about where the sun was.
3 — Desktop editor with layers (best when it has to be exact)
Masking, feathering and manual colour matching still produce the most controlled result, and nothing else lets you leave the rest of the picture literally unchanged. It also takes half an hour and some skill, which is why most people are reading this page instead.
What method one looks like when you add a person to a photo
One portrait, one second image, four merges
Method one, step by step
This is the whole procedure. It takes about a minute.
Step 1 — pick the right portrait
Choose the photo of the missing person whose lighting is closest to the group shot, not the one where they look best. Matching light does more for the result than any other decision you will make here.
Step 2 — upload both
The group photo and the single portrait. Two inputs is the reliable maximum; if several people are missing, do them one at a time or use the family generator built for groups.
Step 3 — name the position out loud
"Place the person from the second image standing at the right-hand end of the group, at the same distance from the camera as the others." Vague placement is the number one cause of a person who is subtly the wrong size.
Step 4 — add the lighting clause
"Match the lighting direction and colour temperature to the group photo." Skip this and you get a technically correct composite that still looks wrong, without being able to say why.
Step 5 — generate a few, then choose
Each run redraws the frame, so results differ. Produce three or four and pick, rather than regenerating the same one hoping it settles.
The prompt that works to add a person to a photo
Copy this and change only the bracketed part.
The prompt
"Add the person from the second image into the first photograph, standing [where], at the same distance from the camera as the other people, matching the lighting direction and colour temperature of the first photograph, keeping their own face, hair and clothing unchanged, photographic realism."
Why "same distance from the camera" is in there
Without it the added person is frequently rendered slightly too large or too small — the single most recognisable tell of an AI composite, and one people notice without being able to name.
Why the likeness clause matters
"Keeping their own face, hair and clothing unchanged" stops the quiet beautification pass: smoother skin, straighter teeth, tidier hair. It is subtle enough that you only notice when a relative points it out.
Four mistakes people make when they add a person to a photo
Every one of these came from our own failed runs.
Using a flash selfie as the source
Direct flash produces a flat, frontal light that exists nowhere in a daylight group photo. The model will try to reconcile it and the face ends up looking pasted in regardless of the prompt.
Leaving the position unstated
The model picks a spot, and picks a different one next run. Every complaint about inconsistency we have seen traces back to this.
Expecting the original photo to survive untouched
The AI route redraws the whole frame. Faces, textures and background detail all shift slightly. When the original must be preserved exactly, that is what method three is for.
Trying to add three people at once
Quality falls off past two inputs, and whichever person you described least is the one that gets dropped or blurred. Sequential runs work; heroic prompts do not.
Adding a person to a photo at a glance
method one, per attempt
~20s
method two, phone cut-out
~1 min
method three, manual masking
~30 min
free credits on signup here
50
Tools for adding a person to a photo
AI Combine Two Images
The generator that runs method one, with the prompt pre-filled.
AI Family Photo Generator
When more than one person is missing from the frame.
AI Photo Restoration
Repair the old group photo before you merge anything into it.
AI Video Upscaler
Bring the finished composite up to print resolution.
How to add a person to a photo — FAQ
The questions people actually ask, answered directly.
How do I add someone into an existing photo?
Upload the existing photo and a portrait of the person, then write one instruction naming exactly where they stand and telling the model to match the original photo's lighting. Those two details — position and light — decide almost the entire result.
What is the easiest way to add a person to a photo?
The AI route, by a wide margin, because it handles the relighting for you. A phone cut-out is faster to learn but leaves a join you can see. Manual masking in a desktop editor gives the best control and takes roughly thirty times as long.
Can I add someone who has passed away?
Yes, and it is one of the most common reasons people look this up. One clear portrait is enough. The result is a new picture rather than a record of an event, and it is worth telling family that plainly when you share it.
How do I add a person to a picture on my iPhone?
The built-in editor will lift a subject from one photo so you can paste it into another in a second app. It is quick and it does not relight anything, so it works best when both photos were taken in similar conditions.
Will the original photo stay exactly the same?
Not with the AI method — it redraws the whole frame, so background detail and other faces shift slightly. If preserving the original pixel-for-pixel matters, use a desktop editor with layers instead, which is the only method that guarantees it.
Why does the person I added look the wrong size?
Because the prompt did not fix their distance from the camera. Adding "at the same distance from the camera as the other people" resolves it in almost every case we tested.
Can I add more than one person?
One at a time works well. Two or more in a single run and quality drops noticeably, with the least-described person losing definition first. For a whole group, use the family photo generator, which is built around multiple inputs.
Is there a free way to do this?
Signing up here gives 50 free credits, enough for several attempts. Phone editors also do a basic version at no cost, with the lighting limitation described above.
Does it work on old scanned photos?
It does, but restore the scan first. Creases and stains are not read as damage to be fixed — they are read as part of the picture, and the model will happily carry that texture across the new person too.
Is it ethical to add someone to a photo?
For family keepsakes, plainly yes. Two lines worth holding: do not present the result as an unaltered record of something that happened, and do not put an identifiable person into a picture they would object to being in.
Add a person to a photo now
Method one runs on this site. Upload the group shot and the portrait, name the position, keep the lighting clause. Free credits on signup.