Blur Photo Background
Portrait-mode style background blur using an on-device segmentation model. Free, private and instant: it runs entirely in your browser.
- Processed locally
- No upload, no wait
- No sign-up
- No watermark
How to use Blur Background
- 1
Open or drop an image.
- 2
Use the controls to get the result you want.
- 3
Download the result. Nothing is uploaded at any point.
Why use Aback Images for blur background
Processed on your device
Your images are decoded, edited and encoded inside this browser tab by Web Workers. Nothing is uploaded to a server.
Instant results
Everything runs locally in a background thread, so there is no upload queue and no waiting on a server.
Batch friendly
Process many images at once and download them individually or as a ZIP.
Every format
Works with JPG, PNG, WebP, AVIF, HEIC, GIF, TIFF, SVG, RAW and more, identified by content rather than file name.
Related tools
AI cut-outs that run on your device, with a touch-up brush, edge softness and new backgrounds.
Cut out the subject with local AI and place it on a colour, gradient or your own photo.
Extract text from screenshots, scans and photos in 12 languages. Tesseract runs in your browser.
Detect faces automatically and blur, pixelate or block them. Adjust or add boxes before saving.
Shrink JPG, PNG, WebP, AVIF and GIF files with MozJPEG, OxiPNG and modern codecs, with a live before/after preview.
Resize to exact pixels, a percentage or a longest side, with Lanczos3 and Magic Kernel resampling. Batch ready.
Strip EXIF, GPS, XMP, IPTC and comments losslessly. Pixels are never re-encoded for JPG, PNG and WebP.
Convert between 34 image formats in bulk: HEIC, WebP, AVIF, JPEG XL, PNG, JPG, GIF, TIFF, SVG, ICO, PSD, RAW and more.
Frequently asked questions
How does background removal work without a server?
A segmentation neural network (U²-Net, or IS-Net in high-detail mode) runs inside your browser with ONNX Runtime and WebAssembly. The model file downloads once from this site and is cached; your photo is never sent anywhere.
Why is the first run slower?
The AI model has to download (about 4.5 MB for fast mode, 44 MB for high detail) and initialise. After that it is cached by your browser and later runs start much faster.
Can I fix mistakes in the cut-out?
Yes. Use the Restore and Erase brushes to paint parts of the subject back in or remove leftovers, then download.
Are my images uploaded anywhere?
No. Processing happens inside your browser with WebAssembly and Web Workers. The site only downloads the app code and codecs; your files never leave your device. You can confirm this with the live upload counter on the tool, or in your browser's network panel.
Can I process many files at once?
Yes. Drop a whole selection or folder. Files are processed in parallel on several worker threads, each result can be downloaded on its own, and everything can be saved as one ZIP.
Is it free? Is there a file limit or watermark?
It is free with no account, no watermark and no daily limit. The only limit is your device's memory: very large images (over roughly 100 megapixels) may be refused to keep the tab from crashing.