Image Diff Checker
Compare two images pixel-by-pixel to detect visual changes and regressions. Highlight differences, adjust tolerance, and export the diff map.
Private by architecture:
All processing is 100% client-side memory. Your data never leaves this browser tab.
1. Source Images
No image loaded
No image loaded
About This Tool
The Image Diff Checker identifies subtle visual regressions, compression artifacts, and layout shifts between two versions of an image.
By comparing individual red, green, blue, and alpha channels on HTML5 Canvas, the tool highlights altered pixels with high-visibility masks and computes exact percentage differences. All analysis executes locally without uploading images to any server.
Key Features
- Pixel-level visual comparison: Detect pixel modifications, color deviations, and subtle design shifts instantly.
- Per-channel tolerance threshold: Filter out imperceptible lossy compression noise by tuning color tolerance from 0 to 64.
- Dimension mismatch handling: Gracefully compares images with different dimensions by expanding the canvas and highlighting offset margins.
- Side-by-side and diff views: Inspect the original, modified, and combined difference mask with one click.
- Export difference image: Download high-visibility PNG difference masks for bug reports and pull request reviews.
Frequently Asked Questions
How does tolerance threshold work?
Tolerance sets the maximum allowable per-channel difference (from 0 to 255) before a pixel is marked as changed. A small threshold (such as 16) ignores imperceptible JPEG compression noise.
What happens if the two images have different dimensions?
The diff engine expands the canvas to the maximum width and height, aligns both images to the top-left origin, and marks missing margin pixels as changed.
Are my images uploaded anywhere?
No. Pixel comparison runs entirely inside your browser memory using Web APIs. No network requests are made.
Can I compare PNG, JPEG, and WebP together?
Yes. The browser decodes any supported raster image format into raw RGBA pixel arrays before comparison.
