Background Removal
AI background removal: a browser-local ONNX model (~60 MB on first load, cached thereafter) works on portraits, products, and pets. Files never leave your browser. Works offline after the first load.
The background-removal engine needs more memory than iOS Safari and in-app browsers (WeChat, QQ, Douyin, Xiaohongshu, etc.) allow per tab, so it can't run here. Please open this page on a desktop browser (Chrome, Edge, Firefox, or Safari on macOS).
How to use
Purpose
AI-powered background remover that runs entirely in your browser — no upload, no daily limit, fully private. Uses the @imgly/background-removal WebAssembly model to segment people, products, pets, and other subjects from any background, producing transparent PNGs with clean hair and edge detail. Unlike remove.bg, Photoroom, Canva, or Adobe Express (all cloud-based), your image never leaves your device — essential for product photography, headshots, and sensitive personal photos.
Steps
- Upload an image with a subject and background (JPG / PNG / WebP)
- First use: the ~30MB AI model downloads from our CDN (one-time, then cached)
- Model auto-segments the foreground subject in 2-5 seconds
- Preview the cut-out, refine edges with a brush tool if needed
- Choose output: transparent PNG / solid color background / a different image as new background
- Download the result
FAQ
- How accurate is it — does hair come out clean?
- We use the @imgly/background-removal model — comparable to remove.bg quality for well-lit subjects. Hair, fur, and translucent edges segment cleanly. Dark hair on dark backgrounds may need manual brush refinement (tool included).
- Is my image uploaded?
- No! This is our biggest differentiator vs remove.bg, Photoroom, Canva — the AI model itself runs in your browser via ONNX Runtime + WebAssembly. Critical for product photography, biometric ID photos, and confidential subjects.
- Why is the first run slow?
- First use downloads ~30MB of model weights from our CDN. Browser caches it; subsequent removals take only 2-5 seconds.
- Can I batch process?
- Yes, drop multiple images. Each runs the on-device model — limit to ~20 at a time depending on your CPU/GPU.
- Can the output feed straight into the ID photo tool?
- Yes. Cut out → export transparent PNG → drop into the ID-photo tool to add the required passport background color and size.
Use cases
- E-commerce product photos on pure white background
- Passport / ID photo background swap to compliant color
- Social avatar with creative background swap
- PowerPoint / Keynote design assets
- Apparel product photography with unified backdrop
Use cases
AI cutouts for products, ID photos, avatars, design assets — model runs on-device, photo never uploaded. Privacy that cloud tools can't match.