ເອັບເຟັກເບຼີພື້ນຫຼັງມືອາຊີບ

Bokeh is the aesthetic blur produced by out-of-focus areas of a photo, typically achieved with a wide camera aperture. Our tool simulates this effect to make your subject stand out against a softly blurred background.

Yes — choose from light, medium, or strong blur to match your visual style. Light blur preserves context; strong blur creates a dramatic portrait look.

Yes — portraits, products, landscapes, group shots. Our AI identifies the main subject and applies the blur only to the background, leaving the foreground crisp.

Yes, and it is the recommended path when you have already done the cutout and just want to add the blurred backdrop. Drop the PNG cutout in and the editor recognises the existing alpha channel — it skips the segmentation pass entirely and lets you load the original unblurred backdrop as a second layer, then apply the blur kernel to that. The output is identical to what you would get by starting from the source photo, just faster (no re-segmentation) and with whatever brush-tool refinement you previously did to the cutout edges preserved. This is the workflow most photographers actually use — do the cutout once carefully, then iterate on the backdrop blur intensity and colour treatment without redoing the segmentation each time. The cross-link to /private-background-remover/ explains the on-device segmentation pass behind the original cutout.

Yes via the bulk-processing entry point. Drop a folder of headshots in, select the same named-radius preset for the whole batch (8 px is the convention for corporate-directory portraits where some context should survive), and the editor runs the cutout-plus-blur pass on every file in sequence. Each output respects the same blur intensity, so the resulting team page or corporate directory reads as a cohesive set with consistent backdrop softening across every member. The /headshots/ vertical documents the LinkedIn-tuned variant of this batch flow with the 800x800 export the LinkedIn API expects, and the bulk-processing rate-limits and concurrency caps documented in our infrastructure pages apply across the run — the pipeline handles a typical 20 to 50 person team page in under five minutes on a recent laptop, with every full-resolution original staying on your device, with no per-file paid quota.