How to Split an Image into Layers Free

"Split an image into layers" means something different for a photo than for a design. A poster made in a design tool was built from layers and an AI can often recover them - the headline, the logo, the shapes. A photo was never layered. The camera recorded one surface, so splitting it means deciding where each object ends, cutting it out, and inventing what was hidden behind it. That last part is why every photo layer tool, ours included, asks you to check the result.

Step by step: photo to layers in the browser

1. Add the photo. Open Layers and drop in a JPG or PNG.
2. Choose Process locally. The subject is cut out, split into one layer per separate object, and each object's edge is re-matted on its own crop so small objects keep clean outlines. The first run downloads the AI once.
3. Review before you accept. You see the original beside the reconstruction built from the new layers. If the background fill behind an object looks wrong, this is the moment to notice it.
4. Edit the layers. Move, resize, reorder, recolor or delete any object. Restore original colors undoes a recolor.
5. Pull out anything that stayed merged. Select an object lets you click it, brush over it, or drag a box around it, preview the selection, and create a new layer from it. Repair shadow marks a cast shadow or reflection so it is cleaned out of the background too, instead of being left behind when the object moves.
6. Export. Download PNG flattens your edited composition; Download PSD keeps every object as a separate named layer.

What separates automatically, and what does not

The automatic split works from the cut-out: every piece of foreground that stands on its own becomes a layer. That is the honest boundary of the feature, and it is worth knowing before you pick a photo.

Two practical limits. Layers are built at a working size of up to 1,024 pixels on the long edge, which suits social posts, slides and mock-ups rather than print. And the tool is in public beta: it shows you the reconstruction side by side with the original precisely so you can catch a bad fill before it reaches a file. If what you need is one full-resolution cut-out rather than several layers, the background remover keeps source resolution.

Why touching objects are the hard part

Finding "the foreground" is a solved problem - it is what every background remover does. Deciding where one foreground object ends and the next begins, when they share an edge, is not. A pair of shoes pressed together has no background pixels between them to follow, so any automatic split has to guess from shape and texture, and a wrong guess cuts a shoe in half or merges a person with the guitar they are holding.

We tested automatic splitting of touching objects against held-out photo sets and it did not beat the simpler rule of one layer per separate piece, so that is what ships. The manual tools exist for exactly this case: a box around one shoe, or a brush stroke over one person, is a few seconds of work and gives an answer you chose rather than one you have to audit.

Other ways to turn a photo into layers

• Photoshop by hand. Select Subject or Object Selection, then Layer via Cut for each object and Content-Aware Fill for the holes. The most control and the most time, and a subscription.
• Upload-based "image to layers" sites. A growing group of single-purpose tools process the image on their servers and return a PSD; free tiers typically cap the number of layers, for example 4 layers or up to 10. Check where your image is stored before using one for client or unreleased work.
• Design tools. Canva Magic Layers is built for flat designs and recovers live text, which a photo tool cannot. Canva notes that photo-realistic images may not separate as cleanly. How the two compare →

Frequently Asked Questions