Blur Images Online
Apply Gaussian blur to any photo with a simple radius slider from 1px (subtle) to 20px (strong). Use it to blur backgrounds for a bokeh effect, obscure faces for privacy, hide sensitive information, or create soft image placeholders. All processing happens in your browser.
Load an image to begin.
How to use
- 1. Drop your image onto the upload zone or tap to select from your device.
- 2. Adjust the blur radius slider — start at 4–6 for moderate blur, 15–20 for strong blur.
- 3. Select output format (JPEG for smallest, PNG for lossless).
- 4. Click "Blur all" to process.
- 5. Download the blurred image.
Frequently asked questions
What is Gaussian blur?
Gaussian blur is a mathematical filter that replaces each pixel's value with a weighted average of its neighbours. The weights follow a Gaussian (bell curve) distribution, producing smooth, natural-looking blur.
What blur radius should I use?
1–3 for subtle softening, 4–8 for visible blur (background effects), 10–15 for strong privacy blur, 15–20 for heavy blur (LQIP placeholders, extreme privacy). Start low and increase until the effect looks right.
Can I blur just part of an image?
This tool applies blur to the full image. For selective area blurring (e.g., just a face or a license plate), you would need a more advanced tool that supports selections.
Is blur effective for hiding sensitive information?
At high radius (15+), blur effectively obscures text, faces, and identifiable details. For maximum security, use a solid colour fill over sensitive areas rather than blur, as some blur algorithms can theoretically be partially reversed.
What is LQIP?
Low-Quality Image Placeholder — a tiny, heavily blurred version of an image shown while the full image loads. Apply maximum blur (20) then compress heavily. The resulting small file can be embedded inline in HTML for instant visual placeholder.
Does blur work on PNG with transparency?
Yes. The blur is applied including the alpha channel. However, canvas blur may cause edge fringing where transparent pixels meet coloured pixels. For transparent images, check the result carefully.