1
Full size
Upload the photo
Upload the full-size original so the repair has every line, texture and grain to continue — the file the camera wrote, rather than a screenshot of it.
One photo in · No masking · No app to install
Remove Person from Photo
Get the memory back the way you remember it, without the four strangers who walked through it.
The background is rebuilt, not smearedNothing to maskThe people you keep come through clean
Before / after
Remove person from photo: a cathedral plaza, emptied
A real run of this tool, and the hardest case on this page — around twenty tourists walking, sitting and photographing across the square. Drag the divider: the woman in red is untouched, the pigeons stay, and the cathedral doorways, the arcade, the bollards and the reflections in the wet cobbles are all rebuilt where the crowd was standing. Picture the shot you queued for, with the square to yourself.
Drag to compare before and after
AI photo tool
What it takes to remove person from photo properly
Deleting someone is the easy part. The work is everything they were standing in front of: the sea wall, the railings, the windows of the terrace, the stripe of wet sand. A tool that only erases leaves a soft grey ghost where they were, and everybody sees it. This one reconstructs the hidden area — continuing lines through, matching grain, holding the same light — so the photograph reads as one taken with the promenade empty.
Image tool
This is the single most requested photo edit there is, and the reason is always the same: the shot is already good. The light was right, both of you were laughing, and there is a man with a red backpack directly over your shoulder. Cropping loses the view. Blurring the background announces itself. Cloning by hand takes an hour and still leaves a repeating pattern in the brickwork.
Here you remove person from photo with one upload and one sentence. The edit runs on GPT Image 2.5 Flare in the browser with free credits to start, the flow opens on image-to-image, and the credit estimate updates before you commit.
Creative engine
How the repair works
Upload the photo, name who goes, and the credit estimate updates before you generate.
1 photo
0 masks
No app
Before: three strangers in frame
A man with a red backpack, a woman with a pushchair, a teenager on a phone — all of them sharp, all of them between the couple and the sea wall.
After: the promenade behind them
The sea wall runs unbroken, the railings keep their spacing, the pastel terrace continues, and the shadows on the pavement belong to nobody who is gone.
Indoors, where the background is furniture
A team shot with a colleague walking past holding a laptop and another at a monitor behind them — both of them clearly in frame, both of them nothing to do with the photo.
The glass wall and the desk put back
The meeting-room glass, the desk edge, the monitor and the plant shelf are continued behind where they stood, at the same flat office light and the same shallow focus. The three colleagues do not change.
Runs in the browser with free credits — nothing to install and nothing to mask. This tool takes people out; add person to photo puts one in from a second upload, and the add a loved one to a photo tool does it with an old album scan.
How to use
How to remove person from photo in three steps
From a crowded frame to a clean one, with the cost shown before you commit.
2
In plain words
Say who goes, not just what to erase
Describe them the way you would to a friend: the man with the red backpack on the left, the two people behind us, everyone except the couple in front. Naming them is faster and more accurate than drawing around them.
3
The step that decides it
Name what should be behind them
This is the step that decides the result. Say what has to continue — the stone wall and the railings, the beach and the sea, the crowd replaced with empty pavement. Get that right and the photo comes back looking like the day you remember, not like a photo that was fixed.
Who it's for
When one person is ruining the frame
A photograph worth keeping, ruined by somebody who does not belong in it. These are the three that come up most when people remove person from photo.
The holiday shot with a crowd in it
Landmarks are busy. You get one photo in front of the thing you travelled to see, and twenty other people are in it.
The photo with an ex in it
A good picture of you that you cannot post, cannot print and will not delete. Take them out and the rest of the photograph is yours again.
Listings, headshots and profiles
A room with someone reflected in the mirror, a team photo with half the office wandering through it, a profile picture with a friend's shoulder in the corner.
Answers
Questions about taking people out of pictures
What people ask before they remove person from photo on Kavel.
Every account starts with free credits and the whole thing runs in the browser, so you can remove person from photo on a real picture of your own today. The credit cost of each run is shown before you press generate, and a render that errors out is refunded automatically.
Upload the picture, describe the person you want gone the way you would to a friend, and say what should be visible behind them. Naming the background is the part most people skip and the part that decides whether the repair holds up at full size.
The underlying image model can, and this page is that capability with the job already set up: the flow opens on image-to-image, the prompt is written for background reconstruction rather than erasure, and the credit cost is on screen before you run it. No prompt engineering and no chat window.
Cutting out is the old workflow — select, delete, then spend the real time repairing the hole. Here there is no selection step at all. You name who goes, and the hidden background is rebuilt in the same pass.
Name what has to stay visible behind it, and the hidden area is reconstructed instead of filled. This is also where it beats erase-and-fill tools outright on overlaps: a shoulder or an arm crossing in front of someone who stays gets rebuilt as that person's own shoulder or arm, because the model understands what it is looking at rather than smudging in nearby colour.
Describe them and say what should be there instead — and give the repair the best odds by picking a shot with structure behind the clutter. Anything structured — walls, railings, pavement, sky, water, foliage — rebuilds cleanly, because the pattern to continue is right there. Give the repair the most detail you can by uploading the full-size original rather than a screenshot.
Kavel pricing — you pay per generation, never per seat
Credits are shared across everything Kavel runs. You see the cost before each run, and a failed job is never charged.
50% OFFYour first purchase is half price
15% OFF - Save $18.8
$8.3/month$9.9/month
$100/year billed yearly
- 9,600 Credits Per Year
- Up to 960 videos
- Up to 1,920 images
- No Queue, Instant Start
- HD, Watermark-Free Downloads
- Commercial Use License
20% OFF - Save $70.8
Popular
$24.0/month
$288/year billed yearly
- 38,400 Credits Per Year
- Up to 3,840 videos
- Up to 7,680 images
- No Queue, Instant Start
- HD, Watermark-Free Downloads
- Commercial Use License
40% OFF
$30.0/month
$360/year billed yearly
- 86,400 Credits Per Year
- Up to 8,640 videos
- Up to 17,280 images
- No Queue, Instant Start
- HD, Watermark-Free Downloads
- Commercial Use License
Free plan
- Get 100 credits a week with check-ins
- 20 Free Credits with login
- Up to 2 videos
- Up to 4 images
- No Card, No Account
- A Free Video With Every Check-in
- Free Engine Every Day
Pay safely and securely with
Payments are processed by Stripe over an encrypted checkout. Your card details are handled by Stripe and are never stored by Kavel.
More tools
Other ways to change who and what is in the shot
The opposite of remove person from photo, plus the other things worth taking out.
Remove person from photo and keep everything else
The picture you already like, with the stranger gone and the wall behind them rebuilt well enough that nobody asks. Your account starts with free credits, so clear the photo you thought the crowd had ruined.