What Is Inpainting and How Does Cleanup.pictures Use It?

Inpainting is a retouching technique that removes unwanted content from a photo and reconstructs the empty area so it blends with its surroundings. Cleanup.pictures applies AI-based inpainting to do this automatically: you upload an image, mark what you want gone, and the tool fills the gap in a few clicks. It suits anyone removing objects, people, text, logos, date stamps, or defects from a picture — photographers, creatives, real estate agents, and e-commerce sellers are the use cases the site names. The free tier handles images up to 720p; higher resolution and a quality refiner sit behind paid plans.

How inpainting works

Traditional inpainting grew out of clone and stamp tools: you copied pixels from one part of the image and pasted them over the problem area, then blended the seams by hand. That approach works but is slow and easy to get wrong — visible smears and lines are a common result.

AI-based inpainting changes the input. Instead of copying pixels, the model reads the surrounding context — texture, lighting, edges, perspective — and generates new pixels that plausibly belong there. The practical difference is that you spend your time selecting what to remove, not repairing the fill.

The workflow is the same in both cases:

1. Input — the original image plus a mask marking the region to remove.
2. Action — the tool fills the masked region using the surrounding image as context.
3. Expected result — the marked object is gone and the area looks like it was never there.

What Cleanup.pictures does with it

The site describes its core function as removing unwanted objects, people, text, and defects from any picture, and frames the whole product around inpainting. Concretely, the page lists these uses:

• Photographers — removing timestamps, removing tourists from holiday photos before printing, cleaning portrait photos for profile pictures, and removing cracks on old photographs.
• Creative agencies — remixing existing photos by replacing parts with their own content, without leaving the creative flow.
• Real estate — depersonalizing and cleaning photos of rooms, flats, houses, or apartments by removing unwanted objects.
• E-commerce — uploading product photos directly and producing cleaner product shots for social media and store listings.
• Text and marks — removing unwanted text, logos, date stamps, or watermarks.

The interface is drag-and-drop: you drop an image above the fold and start for free. The site's own framing is speed — "retouch images in seconds" — and one customer quote speaks to the quality gap versus older tools. Dawn Veltri, Director of Marketing at Raek, describes spending significant time cleaning a picture with similar programs and getting "weird smears and lines," then editing the same photo with Cleanup.pictures in 30 seconds "without the smears and lines."

Why AI inpainting beats clone-tool retouching

The site's own answer to "why is Cleanup.pictures better than other inpainting apps?" is that inpainting used to work like a clone tool, and artificial intelligence now gives much better results. That claim lines up with the mechanism: a clone tool can only reuse pixels that already exist in the image, so it struggles with gradients, repeating patterns, and areas where no clean source patch exists. A generative fill can synthesize what isn't there.

This matters most in the cases the site highlights — removing a person from a busy background, erasing a watermark over textured surfaces, or patching cracks across a face. Those are exactly the situations where a copied patch tends to show.

Resolution, pricing, and access

The pricing table on the page lays out the tiers:

Two things worth noting before you commit. First, the free tier's 720p cap is a real constraint if you're preparing images for print or large displays — check your target output size against it. Second, the "Try Free" link on the page routes to ClipDrop's pricing flow, so the paid path runs through ClipDrop rather than a standalone Cleanup.pictures checkout. If you need inpainting inside your own product, the API is the relevant route and is priced by usage.

When to use it, and when not to

Cleanup.pictures fits tasks where you want a specific object gone and the surrounding area gives the model enough context to rebuild it — a tourist in a landscape, a logo on a wall, a timestamp in a corner. It's a poor fit if you need to add content rather than remove it (that's generative fill, offered in the ClipDrop Pro bundle), or if you need full-resolution output on the free plan.

For a quick test: take one image you'd normally spend ten minutes cloning by hand, run it through the free tier, and compare the fill quality against your manual result. If the 720p limit is the only thing holding you back, that tells you whether Pro is worth it for your volume.