Create a blur background effect online without turning the whole photo soft. Upload one image, let the tool map the subject, then tune the blur radius and edge feather before exporting a real PNG composite.
Source image / 01
Start with one photo where the subject is easy to read.
JPG, PNG, or WebP · 5 MB max · 256–4096 px per side
Result / depth preview
Your result will appear here
The first screen is the workbench. Upload a photo to start.
Blur studies
These visual studies show the decision the tool is built for: keep the subject readable, soften distracting context, and make the result easy to inspect before it leaves the page.
A gentle background blur adds separation around eyes and hair while keeping the original facial pixels as the sharp foreground source.
ReviewInspect hair strands, glasses, shoulders, and high-contrast edges before using the image in a profile or campaign.
A pet photo benefits from depth when the environment matters less than the eyes, muzzle, and silhouette that carry the story.
ReviewCheck ears, whiskers, tail edges, and paws where a subject mask can meet a busy background.
A controlled blur can reduce visual competition around a product without replacing the scene or inventing a new product surface.
ReviewVerify labels, corners, reflections, and the exact product color against the source before publishing.
Definition
A blur background tool separates the foreground subject from the rest of a photo, blurs the background layer, and composites the original subject back over it. The useful result is not a CSS effect on a preview: it is an exported image with a sharp subject, a softened background, and an adjustable edge transition.
The editing problem has two separate parts. First, the image needs a subject mask that follows the person, pet, or product. Second, the background needs a controllable blur that changes depth without changing the subject. Treating those as one generative prompt makes it harder to preserve identity, labels, and small edges.
This page uses a one-time AI segmentation step for the mask. The browser then uses the uploaded image as the foreground pixel source, draws a blurred copy behind it, and applies the mask with an optional feather. That split makes the control predictable: move the sliders to compare different looks without waiting for another model response.
Practical workflow
The sequence is short because the important decision is visible in the workbench. Start with a usable source, review the mask, then export only the version that survives a close look.
Choose a JPG, PNG, or WebP image between 256 and 4096 pixels on each side and no larger than 5 MB. The page checks the actual decoded image, not just the file extension.
Use a source where the subject has visible contrast against its surroundings.
The AI segmentation task maps the foreground once. It costs 10 site credits, and the protected server flow validates the upload, ownership, balance, task state, and failure refund path before processing.
Look at the first result around hair, fur, transparent objects, and product edges.
Adjust Blur radius from 0 to 40 px and Edge feather from 0 to 12 px. The browser redraws the composite locally, then exports the canvas pixels as a PNG rather than capturing the interface.
Compare the original and result before downloading the final composite.
Method selection
Each method changes a different layer. Pick the one that matches the job instead of using a stronger effect to compensate for the wrong edit.
| Method | What changes | Subject pixels | Best use | Main risk |
|---|---|---|---|---|
| Subject-aware blur | Background is blurred behind a mask | Original foreground pixels stay sharp | Portraits, pets, products, and natural depth | Mask halos or missed fine edges |
| Whole-image blur | Every pixel receives the same softening | Subject detail is softened too | Atmosphere, privacy, or intentional softness | Unreadable faces, labels, and edges |
| Background replacement | The original scene is swapped or generated | Subject may be preserved, relit, or reinterpreted | A new setting is more important than the original scene | New lighting, shadows, or context can disagree |
A blur background edit is not a guarantee of perfect segmentation. Fine hair, translucent fabric, reflective edges, and low-contrast subjects still deserve a human review at the final use size.
Use cases
The right amount of blur is a communication choice. Stronger is not automatically better; the background should recede while still explaining where the subject belongs.
Does the face need more separation from a room, street, or busy group scene?
Try this: Begin with a moderate radius and a small feather. Keep eyes, hairline, glasses, jaw, and shoulders under review rather than chasing a dramatic bokeh look.
ReviewCompare the result at thumbnail size and at 100%. The best setting usually makes the face easier to find without making the edge look cut out.
Is the subject expressive but surrounded by toys, furniture, grass, or other visual noise?
Try this: Use a lighter blur when fur or whiskers carry the personality. Increase feather only enough to avoid a hard outline around the silhouette.
ReviewCheck ears, whiskers, tail, paws, and the negative space between legs where the mask may be uncertain.
Would the product read faster if its original environment became quieter?
Try this: Keep the source scene and use a controlled radius. This preserves the product’s material cues while giving the object more visual priority.
ReviewVerify packaging text, logos, reflections, straight lines, and exact color before the image is used commercially.
Commercial comparison
These tools solve overlapping background-editing jobs in different ways. The comparison focuses on the decision that matters here: how much control you retain after the subject is separated. Competitor names are included for clarity; this page does not send users to competitor sites.
Workflow comparison reviewed
| Decision point | Nano Banana Pro | remove.bg | Canva |
|---|---|---|---|
| Primary workflow | Upload once, create a subject mask, tune blur and feather locally, compare, and export a PNG composite. | Background removal with a separate blur-oriented workflow for softening the scene around a subject. | A broader design editor where background focus and layout work can happen in the same project. |
| After the mask is ready | Numeric 0–40 px blur and 0–12 px feather sliders redraw the result without another paid AI task. | Useful when the main goal is a transparent cutout or a quick background treatment. | Useful when the next step is composing the image with text, templates, or other design elements. |
| Foreground preservation | The sharp foreground is drawn from the uploaded source image, not regenerated by the blur control. | The workflow is centered on separation; inspect the current export path for the exact edit behavior you need. | The editor supports many workflows, so the final pixel treatment depends on the selected tool and settings. |
| Best fit | A focused blur-background edit where adjustable depth and a reviewable PNG matter. | Fast background removal when a transparent subject is the main deliverable. | Design production that needs the photo edit alongside a larger canvas or campaign layout. |
| Cost behavior on this page | 10 site credits for the one-time AI mask; local slider changes and exports do not repeat that task. | Plan and export rules depend on the current product path and account. | Plan, feature, and export rules depend on the current editor and account. |
Choose Nano Banana Pro when the desired output is a sharp original subject over a tunable soft background. Choose a cutout-first service when transparency is the deliverable, or a design editor when the blur is one step inside a larger composition.
Operating facts
The public workbench validates these boundaries before a paid task starts. They describe the implemented page, not a promise that every image will produce a perfect mask.
One subject mask costs 10 site credits. Local finishing does not repeat the task.
The browser reads dimensions for immediate feedback. On upload, the server checks the actual image container, size, dimensions, and content moderation result, then stores the image record under the signed-in account. Before generation, the server verifies that the stored reference belongs to that account.
The AI task creates one foreground mask through the existing workbench provider path. The completed PNG is fetched through the authenticated download route, then the browser composites the original upload and mask locally. Task state, output ownership, credits, and refund handling stay in the existing server workflow.
Credits are checked and consumed by the server for the segmentation task. Adjusting Blur radius, Edge feather, or the comparison divider does not create a new task.
Common questions
These answers separate the page’s implemented limits from editorial advice about choosing and reviewing a blur.
The final foreground layer is drawn from the uploaded source image and clipped by the AI subject mask. The background copy is blurred behind it, so the blur control does not regenerate the face, fur, packaging, or other subject details.
Yes. Use Blur radius from 0 to 40 px and Edge feather from 0 to 12 px. Those changes redraw the composite in your browser and do not start another paid segmentation task.
The page accepts JPG, PNG, and WebP images up to 5 MB. Each dimension must be between 256 and 4,096 pixels. JPEG is limited to 1 megapixel, while PNG and WebP are limited to 4 megapixels.
The AI subject-mask task costs 10 site credits on this page. The blur slider, feather slider, comparison control, and PNG export happen locally after the mask is available and do not repeat the 10-credit task.
No image segmentation tool can promise perfect edges for every source. Hair, fur, translucent fabric, reflections, and low-contrast subjects need a close review. Use Edge feather sparingly and compare the result with the original before publishing.
It is a PNG generated from the composite canvas pixels. The download contains the blurred background and masked original foreground, not the surrounding page UI or a CSS-only preview.
Return to the depth lab
Start with a photo you can inspect, create one subject mask, then choose the amount of depth that serves the image. Related tools are available when you need a transparent cutout, generative editing, or a different canvas.