Remove or hide something in a photo
Load a photo, brush over the thing you want gone, then either fill it in from the background around it or hide it. It all happens in this browser tab: the photo is never uploaded and the tool makes no network calls.
Or drag one in. JPG, PNG, WebP or HEIC. Big photos are scaled to 1800px on the long edge so brushing stays smooth.
Brush over the thing you want gone, then press Remove.
What to expect. The fill copies real pixels from your own photo, so it rebuilds sand, sky, sea, grass, hedges, pavement and plain walls convincingly, and it is at its best on small and medium things. It has nothing to invent with, so large removals and busy or structured backgrounds smear. When that happens, hide instead, crop, or use a generative tool. Brush size can also be nudged with the [ and ] keys.
The three ways to remove something from a photo
Every app that takes something out of a picture is doing one of three things. Knowing which one you are using explains most of the disappointment people have with photo editors.
| Method | What it actually does | Good for |
|---|---|---|
| Patch fill | Copies real pixels from elsewhere in the same photo into the gap, choosing patches that continue the structure running into it. The tool above does this. | Small and medium things over background that repeats: sand, sky, sea, grass, hedges, tarmac, plain walls |
| Generative fill | A model invents background that was never in the photo at all. Apple's Clean Up, Photoroom and Picsart work this way. | Large removals, crowded scenes, anything with structure running through it |
| Crop | Cuts the offending part of the frame off. No invention, no smearing, no trace. | Anyone near the edge, which covers most photobombers |
Patch fill is the honest one. It only ever moves pixels you already have, so it cannot put a face or a building in your photo that was not there. That is exactly why it also has a ceiling: no spare background, no fill.
When each one works, and when it does not
The fill in the tool above uses the Criminisi priority order. Rather than filling the gap evenly from the outside in, it looks along the edge of the brushed area for the points where structure is strongest, the line of a plank or the join of hedge and path, and fills those first, then copies the best matching patch from the background nearby. Lines carry on instead of dissolving.
It does well when the thing you are removing is small or medium in the frame, the background behind it repeats, and there is plenty of that same background left and right of it to copy from. A person in the distance on a beach, a bin at the end of a path, a stray branch against the sky, a bag on the sand: these come out clean enough that nobody looks twice.
It struggles when the background has strong structure running through the gap, such as railings, window frames, brickwork or tiled floors. It approximates them, and the eye catches the join. It also struggles when the person overlaps someone you are keeping, because there is no clean edge to work from, and when the subject fills a big part of the frame, because there is simply not enough background left to rebuild it.
Those are not settings problems. No slider makes a patch fill invent a face. When you hit one, hide, crop, or move to a generative tool.
Why the result smears, and what to do instead
Smearing means the fill ran out of matching background and started repeating whatever sat closest to the hole. Most of the time it is the brushing, not the fill, that caused it.
- Work in two or three passes. Take out the body, press Remove, then clean up the leftover feet or bag on their own. Several small gaps fill far better than one big one.
- Brush the shadow too. A person removed but their shadow still lying on the sand looks odder than leaving them in.
- Brush slightly wide. A tight outline leaves a rim of original edge pixels behind, which reads as a cut-out.
- Do not brush across a strong line. If a railing crosses the person, remove up to the railing on each side instead of straight through it.
- Do not brush into another person. If your brush touches someone you are keeping, the fill will happily copy their hair or arm into the gap.
- Check at full size before saving, and hold Compare to see exactly what changed. On a grainy or film-look photo, a filled patch is smoother than everything around it, and that smoothness is the giveaway.
Hiding a stranger's face, responsibly
Often the goal is not a perfect removal at all. You took a photo in a public place, someone else is in it, and you would rather not put a stranger on the internet. For that, hiding beats filling.
It is not only that hiding always works. A blurred or pixelated face is visibly a decision, so anyone looking understands that a person was there and was deliberately covered. A filled-in face quietly pretends nobody was standing there.
Use a coarse setting rather than a delicate one. Light blur and fine pixelation often leave someone recognisable to people who know them, and can sometimes be partly undone. If the point is privacy, make it obvious. The same goes for anything else identifying in the frame: a number plate, a house number, a name badge, a delivery label with an address on it.
One more habit worth keeping: hide before you post, not after. Once a photo is up, the original is out of your hands.
When you need a bigger tool
If the fill smears and hiding is not the look you want, these are the honest alternatives.
- Apple Photos Clean Up, free on recent iPhones. Generative, so it copes with bigger removals, and it is already on the phone. Step by step in our iPhone guide.
- Snapseed Healing, free on any phone. Patch-based like this tool, with the same limits, but handy if you are already editing there.
- Photoroom or Picsart. Cloud generative editors that handle the hard cases, at the price of uploading your photo to a server and usually paying a subscription.
- Crop. Genuinely the right answer more often than people expect, and the only method that leaves nothing to spot.
Whichever you use, do the removal before you apply a look or grain, then edit the whole frame together. A retouched patch under a film grain layer disappears. A filled patch added on top of one never quite does.
Questions people ask
How do I remove a person from a photo for free?
Brush over them in the tool on this page and press Remove. It rebuilds the background from the pixels around them, and works best for someone small in the frame against sand, sky, sea, grass or a plain wall. On a recent iPhone, Apple Photos Clean Up is also free and copes with larger removals.
Does the tool upload my photo?
No. It runs entirely in this browser tab and makes no network requests after the page has loaded, which is also why it needs no account. Your photo stays on your device and the result is saved straight to your downloads as a PNG.
Why does my result look smeared or repeated?
The fill copies real pixels from your own photo, so it needs enough matching background to copy. Busy or structured backgrounds like railings, crowds and brickwork have no clean match, and a large brushed area runs out of source material. Remove in smaller pieces, or switch to blur.
Can it remove someone standing in front of my subject?
Not well. Where the person overlaps someone you are keeping there is no clean edge to work from, so the fill drags parts of your subject into the gap. Crop if they are near the edge, or use a generative editor.
Is blurring a face enough to protect someone's privacy?
Only if you do it coarsely. Light blur and fine pixelation can leave someone recognisable to people who know them and are sometimes partly reversible. Use a strong setting, and cover number plates and addresses while you are there.