What is Background Remover?
A Background Remover computes background remover from the inputs you provide. It applies the standard formula to the values you enter and returns the result instantly, without sending any data to a server. Output is a transparent PNG. The tool.
Background Remover
Drop an image and get a transparent PNG with the background erased. AI runs in your browser. Files never leave your computer.
TLDR
Drop a photo and a U2-Net-style model running in your browser via WebAssembly produces a clean cutout with transparent background. The first run downloads the model (about 40 MB); after that it stays cached. Best on portraits and product shots; struggles on hair, glass, and fine fur.
Drop an image here or click to choose. JPG, PNG, WebP. Stays in your browser.
How to use this tool
- Wait for model load. The first visit downloads the AI model (about 40 MB). The status line shows when it's ready.
- Drop your image. Drag a JPG, PNG, or WebP photo onto the drop zone, or click to choose.
- Wait for processing. The model removes the background. Takes 5-30 seconds depending on image size and device speed.
- Download PNG. Click Download PNG to save the cutout with transparent background. Drop it into Canva, Figma, or PowerPoint.
About this tool + how it works
This tool runs 100% in your browser - the libraries load from a public CDN and the math runs on your device. Nothing is uploaded to a server. The underlying logic is:
1. The browser loads @imgly/background-removal (WASM + ONNX model, around 40 MB on first run). 2. The model runs a foreground/background segmentation on each pixel. 3. The original image is composited with the predicted alpha mask. 4. Output is a PNG with transparency you can drop into Photoshop, Canva, Figma, or a presentation.
You can verify by opening the browser developer tools and watching the Network tab; you'll see no requests fired during normal use beyond the initial page and library load.
Real-world scenarios where this tool helps
E-commerce product photos
You shot the product on a desk; the marketplace wants pure white. Run it through, then composite onto white. Faster than booking a photographer.
Profile picture cleanup
LinkedIn or company directory needs a clean headshot. Remove the messy background, drop in a solid color or company brand color.
Slide deck assets
Presentations look amateur when speaker headshots have visible backgrounds. Cut them out so they sit cleanly on the slide layout.
Social media graphics
Drop product cutouts into Instagram and TikTok templates. The transparent PNG composites without halo or fringe.
What this tool does
- Removes the background from a foreground subject (person, animal, product) and outputs a transparent PNG.
- Runs entirely in your browser using WebAssembly - no server upload, no API key, no rate limit.
- Caches the model after first download so subsequent images are faster.
- Handles JPG, PNG, and WebP input.
- Preserves the original resolution of the input image.
What it does NOT do
- Doesn't perfectly cut out fine hair, fur, or transparent objects (glass, smoke). For those you may need manual touch-up in Photoshop.
- Doesn't work well on subjects that blend into the background (low contrast).
- Doesn't add a new background - only removes the existing one. Composite the result onto your chosen background separately.
- Doesn't process video. Frame by frame is possible but slow.
- Doesn't run on very old devices. The WASM model needs a modern browser and decent RAM (2 GB+ free).
Common mistakes and pitfalls
- Refreshing the page mid-process and losing the cached model. Wait for the model load to finish before processing.
- Expecting magic on poor source images. Garbage in = garbage out: blurry, dark, or low-contrast photos give blurry, dark cutouts.
- Trying to remove backgrounds from group shots and getting fused subjects. The model is best at single subjects.
- Forgetting that the output is PNG, not JPG. JPG does not support transparency - never re-save the cutout as JPG or the background turns black.
- Comparing to high-end paid services. Local AI is impressive but not as polished as remove.bg's premium tier; expect 80-90% quality for free.
Frequently asked questions
Is this background remover really free?
Yes. Free forever, no signup, no watermark, no daily limit. The model runs in your browser, so there's nothing to charge for.
Where does my photo go?
Nowhere. The image is processed by an AI model running in your browser via WebAssembly. No upload, no logging, no cloud.
Why is the first run slow?
The first time you use the tool, the browser downloads the segmentation model (about 40 MB). After that the model is cached and processing takes 5-15 seconds per image.
How accurate is the cutout?
Excellent on clear subjects with high contrast. Good on most portraits and product shots. Weak on fine hair, glass, smoke, or low-contrast scenes. Compare to remove.bg or Photoshop's Select Subject for the trickier images.
What format is the output?
PNG with alpha transparency. You can drop it into Photoshop, Figma, Canva, PowerPoint, or any tool that supports transparent images.
Can I use it on commercial product photos?
Yes. The output is yours to use however you want. The model itself is open-source and the tool runs locally; no usage license is enforced.
How big can the input be?
Up to about 4000x4000 pixels comfortably. Larger images may run out of memory in the browser. Resize first if you have very high-resolution shots.
Why does it sometimes erase part of the subject?
The segmentation model can confuse foreground and background when colors blend (white shirt on white wall, dark hair on dark backdrop). Crop tighter or shoot against a contrasting background for best results.
Does it work on phones?
Yes on most modern phones with at least 4 GB RAM. The first model download takes longer on slower connections.
Can I run it offline?
Yes after the first run. Once the model is cached, you can disconnect from the internet and continue removing backgrounds.