How We Process Images Locally
nobackground runs AI background removal directly in your browser using WebAssembly and ONNX Runtime Web. Your images are not uploaded to our servers for processing.
What happens first
On first use, your browser downloads the AI model files and caches them. This initial step can take longer depending on network speed.
How processing works after that
After model cache is ready, background removal runs locally on your device. This improves privacy and usually reduces repeat latency.
Low-memory fallback
On lower-end devices, we can automatically fallback to a lighter model to improve reliability and avoid processing failures.
Why on-device processing matters for privacy
Because the AI model runs inside your browser's own sandbox, your original photo and the version with the background removed both stay on your device from start to finish. No image data is transmitted to nobackground's servers, stored in a database, or logged anywhere we control, which is a meaningfully different privacy posture than tools that upload your file to a remote server for processing.
Speed and performance after the first use
The one-time model download is the only step that depends on your internet connection. Every processing run after that is limited only by your device's own CPU or GPU, which is why repeat use on the same device tends to feel noticeably faster than the very first image you process. Closing and reopening the tab does not require re-downloading the model, since modern browsers cache it locally.
Why this matters for sensitive or personal images
People use nobackground for photos they would rather not upload anywhere, such as passport and ID photos, medical or insurance documents, unreleased product shots, or private family pictures. Because processing never leaves the browser, there is no server-side copy of the image to worry about, no third-party API in the middle, and nothing to delete afterward, since nothing was stored in the first place.
Local processing questions
Does nobackground ever see or store my images?
No. Images are processed entirely within your browser's WebAssembly runtime. We do not receive, store, or have access to the images you process.
Will this work if I lose my internet connection?
The AI model needs to download once over an internet connection. After that initial download is cached by your browser, background removal itself can continue to work offline on the same device and browser.
Why does the first image take longer than the rest?
The first run includes downloading and initializing the AI model files, which is a one-time cost. Every image after that only needs the actual processing step, which is much faster.