Upload one photo to this AI photo editor online free and the AI splits it into separate subject, background and object layers. Swap a sky, restyle a backdrop or tidy a product shot by editing those layers in Photopea, Photoshop or any other PSD editor. Export a layered PSD or a PNG ZIP with no watermark.
Upload Image
PSD + PNG ZIPNo Prompt NeededJPG, PNG or WebP up to 12 MB. Sign in to upload.
Click or drag an image here
Sign in to upload your image
Layer count
Free accounts can use up to 4 layers; subscribers get up to 8.
Example Split
Layer Previews
Check every layer before you export
What comes back
One upload returns a stack of transparent layers and an export that opens in the editor you already use.
People, products and pets come back as transparent PNG layers, already cut away from their surroundings. Move, scale or recolor them without tracing an outline first, although soft edges such as hair can still need a touch-up in your editor.
The background arrives as its own layer, and the area that sat behind foreground objects is usually filled in. Results vary from photo to photo, so check that patch before moving a subject far from where it stood.
Props, signs, text and cast shadows often land on layers of their own. Hide one to see the photo without it, or keep a shadow layer to ground your subject on a new background.
Download a layered PSD that opens in Photoshop, Photopea, GIMP and Affinity, or a ZIP of PNG layers for Canva, Figma and slide decks. Neither export carries a watermark.
Before and after
This AI photo editor takes care of the slow part: cutting subjects, background and objects apart. The creative edits happen afterwards in Photoshop, Photopea or another PSD editor, and each example below notes which part the layers actually help with.
Layered PSD or PNG ZIP, no watermark
Portrait & Social
Blemish cleanup like this is hand retouching with a healing brush in your PSD editor. Splitting first puts the person on their own layer, so the backdrop stays untouched while you work and can be restyled for each post.
Product & E-commerce
Keys, books and snack bags that land on their own layers can simply be hidden in your editor, leaving just the product. Anything that shares a layer with the tabletop still needs a quick patch by hand.
Business & LinkedIn
Place an office or studio backdrop under the subject layer and match the color so it sits naturally. Bigger changes, like the suit in this example, mean compositing clothing by hand and take real editing skill.
AI photo editor use cases
The AI does the separating, and the edits below happen in the PSD editor you already use.
Set the product layer on a clean color or a lifestyle scene, and keep its shadow layer when one comes back so the item still sits naturally.
Blur, darken or replace the background behind a person without drawing a mask first. Skin and detail retouching remains manual work in your editor.
Reuse one subject layer across square, story and banner crops, giving each post its own background without cutting the subject out again.
Place model shots on branded backgrounds and assemble campaign composites from separated layers instead of cutting every look out by hand.
Swap a flat gray sky above an exterior or hide loose items when they land on their own layers. Restaging a room with new furniture is still manual compositing in your editor.
Rearrange elements, test new backdrops and try color treatments one layer at a time before committing to a final composition.
How it works
Upload, split and export here, then make the actual edits in the PSD editor you already use.
step 1
Sign in and upload a JPG, PNG or WebP photo of up to 12 MB.
step 2
Pick 2 to 8 layers: up to 4 on a free account, up to 8 with a subscription. Each run costs 10 credits and usually takes 1 to 5 minutes.
step 3
Preview every layer, export a layered PSD or a PNG ZIP, then make your edits in a PSD editor such as Photopea or Photoshop.
Before you upload
Straight answers on quality, cost and speed before you upload anything.
The model, Qwen Image Layered run through fal, decides what belongs together and returns each part as a transparent RGBA PNG layer: typically the subjects, the background, and shadows or objects.
A run costs 10 credits no matter how many layers you ask for. If it fails, those credits return to your account automatically, and finished exports never carry a watermark.
A low count gives broad parts such as subject and background, while a higher count tends to pull props, text and shadows apart. Free accounts go up to 4 layers and subscribers up to 8.
Expect 1 to 5 minutes per run. Layers are sized to the model's working resolution, so they can come back smaller than your upload. That is fine for web and social work; plan print jobs with it in mind.
Beyond the AI photo editor
Related tools and guides for splitting, cleaning up and opening your photos.
FAQ
What the AI photo editor does and does not do, what a run costs, and where the actual edits happen.
It splits one photo into separate transparent layers. An AI model, Qwen Image Layered run through fal, studies the picture and returns the subjects, the background and often shadows or objects as RGBA PNG layers. You preview them, export a layered PSD or a ZIP of PNGs, and make the actual edits in a PSD editor.
You can start with a free account. Every run costs 10 credits whatever the layer count; free accounts can choose up to 4 layers and subscribers up to 8. Current credit packs and plans are listed on the pricing page.
No. The tool does not take text prompts and does not generate images from text. The AI decides how to split the photo by itself, and creative changes such as a new sky, a different backdrop or a color grade are made by you afterwards in a PSD editor.
Split the photo, open the PSD, place your new sky or backdrop just below the subject layer and hide the old background. If the sky shares a layer with the landscape, you will still need to mask it by hand. Finish by matching color and light so the swap looks natural.
Not with one click. When an object lands on its own layer you can hide or delete it, and the background layer usually has the area behind it filled in. Results vary, so zoom in and clean up any leftovers with a clone or healing tool in your editor.
No. There is no upscaling, sharpening, restoration, colorizing or filter feature. The AI only splits a photo into layers, and any repair or color work happens afterwards in your own editor.
Photoshop, Photopea, GIMP and Affinity all open it with each layer listed separately. Photopea runs free in a browser tab, which makes it the quickest option if you have no desktop editor. Tools that cannot read PSD files can use the PNG ZIP instead.
JPG, PNG and WebP images up to 12 MB, one photo per run. You need to be signed in to upload.
Most runs finish in 1 to 5 minutes. Layers come out at the model's working resolution, not your original size: a 1024×1024 upload returned 640×640 layers. That suits web and social images better than large prints.
The 10 credits for that run are refunded to your account automatically, so you can try a different layer count or a clearer photo without paying twice.
Splitting a photo does not change who owns it. If you hold the rights to the image you upload, you hold the same rights to the layers made from it; the tool grants no license you did not already have.
Uploads are used to produce your layers and exports. Read the Privacy Policy for how your data is handled and how to ask for it to be deleted, especially before uploading client work or anything confidential.
Sign in, upload a JPG, PNG or WebP photo and the AI photo editor returns transparent layers, typically the subject, background and objects, ready to export as a layered PSD or PNG ZIP with no watermark.
