Background Remover
Remove image backgrounds with a segmentation model that processes the selected image in your browser after its model assets load. Export a transparent cutout or a flat color fill.
How the AI cutout works in your browser
Background Remover uses an AI model inside the browser to isolate the foreground subject from a photo or graphic. The Transformers module loads from jsDelivr, while model and configuration files for onnx-community/BEN2-ONNX load from Hugging Face infrastructure and may be fetched again when a cached copy is unavailable. The cutout itself is then computed on your device.
This is useful for product cutouts, branded content, headshots, and mockups. The tool code does not intentionally upload the selected image for segmentation, although ordinary delivery, consent, and diagnostic requests can still occur. There are no per-image fees or added watermarks.
When the subject matters more than the scene
Background removal is the right step when the subject of the image matters more than its surroundings. Product photography, passport-style headshots, social media profile pictures, and catalog imagery all benefit from isolating the subject before placing it on a new background or a clean white canvas. It is also the first step in creating composite images, collages, or layered designs where multiple elements need to sit together.
There are also practical workflow reasons to remove backgrounds. Many marketplaces require white-background product photos. Job portals expect clean headshots. Presentation slides look more polished when distracting backgrounds are removed from photos before placing them on branded templates. If you are deciding where this step belongs, follow the background-removal photo editing workflow before final retouching, resizing, and export.
Choosing an export format for the cutout
PNG stores the resulting cutout and transparency without a lossy export, which makes it a practical default for review and further editing. The segmentation step has already estimated and changed edge pixels, so inspect hair, fur, and soft boundaries before use. WebP also supports transparency where the destination accepts it. JPG has no transparency channel, so choose it only when you apply a solid fill color.
How To Remove a Background Here
- Select a photo or graphic from your device using the file picker or drag and drop.
- Choose your output format: PNG for transparency, WebP for smaller transparent files, or JPG with a solid fill color.
- Optionally check the solid background box and pick a fill color if you want a non-transparent result.
- Click Remove Background and wait while the downloaded model processes the selected image in your browser.
Mistakes That Ruin a Cutout
Choosing JPG output and expecting transparency. JPG does not support alpha channels, so always use PNG or WebP when you need a transparent cutout.
Uploading very low-contrast images where the subject blends into the background. The AI performs best when there is a clear visual separation between foreground and background.
Skipping the fill color step when exporting JPG. Without a chosen fill color, transparent areas default to white, which may not match your design needs.
Processing extremely large images on low-end devices. If your phone or older laptop struggles, resize the image first using the Resize tool before running background removal.
How your file is processed
Background Remover loads its Transformers module from jsDelivr and model or configuration files for onnx-community/BEN2-ONNX from Hugging Face infrastructure. Those hosts receive ordinary connection data and the files may be requested again if the browser cache is unavailable. The cutout is computed on your device; the tool code does not intentionally upload the selected image for segmentation.
Frequently Asked Questions
How clean are the edges around fine detail like hair or fur?
The cutout is produced by a segmentation model that handles clear, well-lit subjects against contrasting backgrounds well, giving crisp edges on solid shapes. Wispy hair, fur, and motion-blurred strands are the hard case: a few stray pixels may stay or drop. Even lighting and a background that differs in tone from the subject give the model the best chance.
Why is the very first cutout slow, then later ones quick?
Processing may fetch the segmentation model and configuration files from Hugging Face after the Transformers module loads from jsDelivr. A browser can reuse cached files, but it may request them again if that cache is unavailable or cleared. The tool code processes the selected photo on your device and does not intentionally upload it for segmentation; ordinary page, consent, or diagnostic requests may still occur.
Can I export the cutout with a real transparent background?
Yes. Choose PNG (or WebP) as the output and the removed area is saved as genuine transparency, so the subject drops cleanly onto any design, slide, or listing behind it. Avoid JPG for this, since JPG has no transparency channel and would fill the empty area with a solid color instead of leaving it see-through.
Related Editing Tools
Background removal is often the first step: isolate the subject, then add a new background, adjust colors, or resize for the final destination.
Add Background
Place your cutout on a new solid or gradient background after removing the original.
Open Add BackgroundBlur / Pixelate
Blur or pixelate parts of an image instead of removing the entire background.
Open Blur / PixelateWatermark Image
Add a watermark to your cutout or final image before sharing publicly.
Open Watermark ImageCompress Image
Reduce the file size of your exported cutout before uploading to a website or marketplace.
Open Compress ImageResize Image
Match exact dimensions required by marketplaces, social platforms, or print services.
Open Resize ImageImage Format Converter
Convert your cutout between PNG, WebP, JPG, and other formats as needed.
Open Image Format Converter