Image diff.
Compare pixel by pixel.

See what changed between two images with a difference mask and an adjustable overlay.

Compare images

Reference image

The first image is your reference.

No image chosen.

Comparison image

Choose the image to compare against it.

No image chosen.

Choose two images, then compare to see their decoded previews. PNG, JPEG, WebP, GIF, TIFF, AVIF, SVG and other browser-decodable images. One still frame per file.

Comparison settings

0 catches every compared channel change. A pixel changes only above this value (0–255).

RGBA compares premultiplied color and alpha. On white compares their white-background appearance.

Padding uses the larger width and height. Images are never resized.

Choose a reference and a comparison image.

How to compare

  1. Choose a reference image and a comparison image.
  2. Set tolerance, alpha mode and alignment, then choose Compare images.
  3. Inspect the mask or overlay. Download the selected PNG and a report of the exact settings and counts.

All decoding, comparison and PNG export run in your browser. Files stay on your device and are never uploaded or stored by this tool. Reset releases the images and results.

Questions & details

What counts as a changed pixel?

We compare oriented, decoded 8-bit sRGB pixels. A pixel is changed when its largest channel difference is strictly greater than the tolerance. No anti-aliasing differences are excluded. A zero count means no pixels exceeded your tolerance; it does not mean the source files are byte-identical or perceptually identical.

How are color and transparency compared?

RGBA mode compares red, green and blue multiplied by alpha, plus alpha itself. Each compared color channel is rounded to 8 bits. This ignores invisible RGB in fully transparent pixels. On white instead compares rounded RGB composited on white, so different transparency can look identical. Browser decoding and color conversion may vary; metadata, HDR precision and animation are not compared.

What if dimensions are different?

Equal dimensions are required by default. Top-left and Center place both images on a canvas using the larger width and larger height, with transparent padding. Center rounds half-pixel offsets down. There is no resizing, cropping or automatic content alignment. Padding participates in the count, including areas outside both images.

What do the downloads contain?

The mask is opaque magenta at changed pixels and transparent elsewhere. Overlay is a blend of the aligned images: 0% is the reference, 100% the comparison. RGBA uses a premultiplied-alpha blend; On white produces an opaque blend. PNG retains no source metadata. JSON contains dimensions, offsets, counts and settings, without your file names.

Which images and browsers work?

Use images your browser can decode, with additional TIFF, AVIF and SVG decoder support. Animated or multipage files use one decoded still frame. SVG is rasterized with the static image conversion; external resources are not fetched. Use a current browser that can create PNG images. Actual browser canvas and memory limits still apply; Cancel stops processing if a large comparison takes too long.