The object removal tool we built because we were tired of messy workarounds
Last summer, a user emailed us a photo of her daughter at a seaside shoot. A bin sat in the background, dead centre, ruining an otherwise perfect frame. She'd already tried three other apps. All of them either left ghosting artefacts or demanded she subscribe to 'premium removal credits'. She asked: why is this so hard?
Why we chose PatchMatch over the obvious route
When we started building Cleanr, object removal felt non-negotiable. But every approach came with trade-offs. Generative fill was tempting - draw a box, let it synthesise content. That works for minor tweaks. But ask it to remove a lamppost from a textured stone wall, and you're often stuck with an uncanny blotch that looks nothing like the surrounding pixels.
We landed on PatchMatch because it does something simpler and, frankly, smarter for real photos: it studies the pixels around the object you want to remove, finds matching patches elsewhere in the image, and blends them in. No hallucination. No weird synthesis. Just intelligent borrowing from what's already there. A bin disappears because the water behind it is already in the image. A stray person vanishes because the grass they were standing on is captured right next to them.
It's less flashy than some tools. But it's more honest. And that matters when you've got five minutes before you post.
The brush, the preview, and the moment it clicked
The first week after launch, someone asked if they could adjust the brush size. Seemed obvious in hindsight. We'd locked it at a default that worked fine for 70% of use cases and felt clunky for everything else. Within two days, we'd added size and hardness controls, a live preview that updates as you paint, and a simple undo button for mistakes.
That interaction loop matters more than people realise. You brush over the object, you see the result instantly, you adjust if needed. No waiting. No submitting to a server farm and checking back in 30 seconds. Just you, the photo, and the tool responding in real time. One creator told us it cut her cleanup workflow from 20 minutes per photo down to three.
The speed isn't magic. It's just local processing, a decent implementation, and respecting your time.
When it works, and when you might need to think differently
Object removal shines when you're working with distinct objects against textured or repetitive backgrounds. A person in a crowd, a stray branch across a landscape, graffiti on a brick wall, a parked car in a street shot. The tool has clean edges to reference and patches to borrow from everywhere.
It's less reliable when the object is woven into the background. A person standing in front of a unique mural has no 'clean' version of that mural to reference. A tree limb that branches into a complex canopy? The algorithm struggles because it's trying to invent detail it genuinely doesn't have. That's not a weakness of PatchMatch; it's just the boundary of what content-aware removal can do without fabrication.
We've learned that the best results come when you're deliberate: remove clutter, not complexity. And if you've got a photo where removal feels impossible, try object removal in Cleanr on the Plus or Pro tier first. You get three removals free every day to test. If it works, brilliant. If it doesn't, you haven't wasted credits or money.
Why we kept it separate from generative fill
On the Pro and AI Pro tiers, we offer generative fill as well. It's a different beast: you brush a region, and the app synthesises new content to fill it. It's powerful for creative edits. But we kept it distinct from object removal on purpose.
Object removal is about restoration. You're fixing a photo that was already good. Generative fill is about creation. You're adding something that wasn't there. Lumping them together under one confusing 'remove' button felt dishonest. Someone deletes a photobomb with object removal and gets an artefact, they're frustrated. Someone uses generative fill to fill a gap and gets a watercolour effect instead of photorealistic content, and they understand the tool differently because we've been clear about what it does.
Clarity beats feature count every time.
The real reason we built this
Here's the thing nobody talks about: object removal is rarely the feature that makes someone download an app. It's the feature that makes them stay. Someone opens Cleanr to quickly enhance a photo, then notices a stray object they'd normally live with, tries the removal tool, and finds it works. That moment - when a tool feels reliable and fast - shapes whether they come back.
We built object removal because it's part of being a complete photo studio. You shouldn't need four apps to clean up a photo, remove a blemish, replace the sky, and export it. And you shouldn't need a subscription just to remove one unwanted object from an image you've already taken.
The bin in that original email? The user sent us a follow-up photo six months later. Cleaned up, sky replaced, and framed. She's been with us since.
Object removal sounds simple until you actually need it. Then it's either a time-saver or a frustration. What's the difference between the two?