Group & family photos
Keep several visible people together in one foreground mask instead of splitting the photo into separate files.
Browser-native cutout studioRemove the background.
Remove the background.
Keep everyone.
Clean cutouts for portraits, product shots and photos with multiple people. No crowded editor, no account, no upload queue — just the tool you came for.
Multiple subjectsTransparent PNGPhoto stays on your device
ZERO BG ENGINE
Preparing the AI engine
Downloading model files for local processing.
Preparing…
First load only. Photos stay on your device.
Drop an image here or click to upload
JPG, PNG or WebP · up to 20 MB
No account or watermarkPhoto processed in your browserFocused on the job
One tool. Different kinds of photos.
Start with the photo you already have and create one clean foreground cutout for the next step in your workflow.
Portraits & profiles
Separate a person from a room, wall or outdoor scene for a profile image or presentation.
Product photography
Create transparent product cutouts for store listings, catalog layouts and quick mockups.
Team & event images
Prepare a shared cutout for event graphics, team pages, announcements and internal decks.
Creator assets
Reuse a person cutout across thumbnails, social graphics, slides and other visual content.
Transparent PNGs
Download an image with an alpha channel so your next design can show its own background.
A clean cutout, without an editing-suite detour.
Upload, inspect the result and export. The workspace keeps the controls close to the image instead of surrounding it with extra steps.
Local image processing
The selected photo is processed in your browser. The AI runtime and model files are fetched separately when needed.
Before / after comparison
Switch between the source and the cutout, or use Split view to check the boundary in context.
Zoom into details
Inspect hair, shoulders, small accessories and product edges before saving the result.
Transparent PNG output
Download a PNG with transparency preserved and the source image dimensions retained.
No account or queue
Open the studio directly and work without a sign-up flow or server-side processing queue.
One image, one workspace
Keep the task focused: process a photo, review the output and then start another image.
Questions, answered
ZeroBG FAQ
What the tool does, how the local engine works, and what to expect from your image.
Does ZeroBG upload my photo?
No. The selected image is processed inside your browser tab. The browser downloads the AI runtime and model files from their public distribution hosts on first use, but your photo is not sent to those hosts by this app.
Can it keep multiple people in one image?
That is a main use case. BiRefNet Lite predicts a foreground mask for the full image, so multiple visible people can remain together. Results can vary if people overlap heavily, blend into the background or are partly outside the frame.
Is the background removal free?
Yes. ZeroBG has no account, credit counter or watermark. Processing speed depends on the device and the browser's GPU or WebAssembly support.
Will the exported PNG keep the original dimensions?
The output canvas uses the source image's pixel dimensions. Fine edge quality is limited by the model's predicted mask and the quality of the original photo.
Why does the first run take longer?
When the workspace first comes into view, the browser downloads the AI runtime and model weights, which are roughly 94 MB for the pinned FP16 model before caching and other assets. Later visits can reuse cached files, depending on browser storage settings.
Which file types are supported?
The current browser workspace accepts JPG, PNG and WebP images up to 20 MB. Convert HEIC/HEIF to JPG or PNG first.