Two-person photo, a real split: the flat image and the 2 named layers it became.
Two-person photo
2 layers in 46 s · Instruction: “Separate the woman on the right”
Put Any Object on Its Own Layer — Just Name It
Separating one object into a layer is the LayerGrab mode where you describe the object and get it back as a transparent layer, with the picture behind it filled in. No selection, no mask.
Image
Leave empty to split every element.
Sample split, free to explore: Two-person photo
Asked for: “Separate the woman on the right”
2 layers
Split in 46s
- Background (filled)
Real splits, not mockups
Every example on this page is a real result: the original, the filled background and each named layer, exactly as the tool returned them.
Coffee shop poster, a real split: the flat image and the 11 named layers it became.
Coffee shop poster
11 layers in 100 s
Headphones product banner, a real split: the flat image and the 8 named layers it became.
Headphones product banner
8 layers in 66 s
Skincare Instagram post, a real split: the flat image and the 8 named layers it became.
Skincare Instagram post
8 layers in 80 s
AI-generated desk illustration, a real split: the flat image and the 10 named layers it became.
AI-generated desk illustration
10 layers in 100 s
Anime room illustration, a real split: the flat image and the 7 named layers it became.
Anime room illustration
7 layers in 63 s · Instruction: “girl with dark blue hair holding an open book, orange sleeping cat, monstera plant in a white pot, stack of three books, mug of coffee on a wooden tray, paper lantern lamp on a stand”
Growth report slide, a real split: the flat image and the 12 named layers it became.
Growth report slide
12 layers in 97 s
Sportswear flat-lay photo, a real split: the flat image and the 16 named layers it became.
Sportswear flat-lay photo
16 layers in 148 s
How to separate an object into its own layer
1
Upload the image
A photo, poster, product shot or AI image as PNG, JPG or WebP, up to 20 MB.
2
Name the object
In "What to separate", describe it the way you would to a colleague: "the woman on the right", "only the 20% OFF badge", "the headphones". No selection or masking.
3
Split
The AI returns that object by itself and repaints the background behind it. It takes 60–100 seconds.
4
Use the layer
Copy it straight into Figma, Photoshop or Canva, download it as PNG, or export the object and the filled background together as a PSD.
Separate an object into a layer
One object out, the background repaired
Separating an object into a layer is the LayerGrab mode where you name one thing in a flat image, such as "the woman on the right" or "the price tag", and get back just that object as a transparent layer plus the picture without it, with the gap filled in.
It is the same model as the full image to layers split, pointed at one thing. You pay for what comes back, so asking for one object costs less than splitting everything.
Name it instead of selecting it
The instruction for this photo was a plain sentence: “Separate the woman on the right”. The result is exactly one layer, named “Right-side young woman”, placed where she stood.
Position words (left, right, top, in front), colours and what the thing is all help the model find the right object.
Pick one element out of a design
On a product banner, ask for “only the headphones and the price tag” and you get those two layers, ready to drop into a new layout, while the rest of the banner stays untouched in the background layer.
Doing it by hand in Photoshop
Photoshop has no single command for this. The usual route is three steps: select the object (Select > Subject or the Object Selection tool), copy it to a new layer with Layer via Copy (Ctrl/Cmd+J), then remove it from the original and fill the gap with Content-Aware Fill.
That is quick for a product on a white background. It gets slow when the object overlaps other things or when what is behind it has to be invented, which is the part LayerGrab does for you. If you work in Photoshop all day, the Photoshop plugin runs the same split inside your document.
When not to use this
If you only want the object on a transparent background and do not care about the picture behind it, a background remover is faster and usually free.
The layer is pixels: text stays an image, not editable type, and a shadow the object casts on the ground is not carried with it. For a whole design where every element should become a layer, use the full image layer separator or export a layered PSD.
Frequently asked questions
How do I make an object its own layer in Photoshop?
Select the object (Select > Subject, or the Object Selection tool), then choose Layer > New > Layer via Copy (Ctrl+J on Windows, Cmd+J on Mac). The object is copied to a new layer above the original. If you use Layer via Cut instead, the object moves to the new layer but leaves a transparent hole in the original, which you then have to fill, for example with Edit > Content-Aware Fill.
How is this different from Layer via Copy?
Layer via Copy needs a good selection first and leaves the original object in the background layer, so moving the new layer shows a duplicate underneath. Here you only name the object, and you get two things back: the object alone, and the background with the object removed and the area behind it painted in.
Can I separate more than one object?
Yes. Name them together, for example "the headline and the product", and each one comes back as a layer of its own. Leave the box empty to separate every element in the image.
Does it work on photos of people?
Yes. The sample above is a two-person photo split with the instruction "Separate the woman on the right": she comes back as one layer and the background behind her is filled in. Fine hair and very busy backgrounds can leave a softer edge than a careful manual mask.
What does it cost?
Credits are charged per layer you get back, and the filled background counts as one. Separating one object is two layers. New accounts in supported regions get 30 free credits, which covers a split of up to 3 layers.
Last updated 2026-09-29