Texloom
Pricing
Sign in
Launch Studio
Free Tools
Image Diff Compare
Free · instant · no upload

Image Comparison Tool

Drop two images and spot the differences — slider overlay, side-by-side, and pixel-diff heatmap with a change-percentage score.

Instant diff Three view modes Change score
Unlimited · 100% free
Learn more

How image comparison works — pixel diff, explained

Image comparison at its simplest: subtract pixel A from pixel B, take the absolute value, and you get a per-pixel difference map. Average that map and you get a difference score — a single number telling you how much two images differ. Zero means byte-identical. 1.0 means maximum possible difference (white vs black). Most real-world compares land between 0.001 (same image saved twice) and 0.3 (substantively different images).

This tool normalises both inputs to 512×512 so images of different sizes can still be compared. It then produces three views: a drag-to-reveal slider, a side-by-side grid, and a bright-red heatmap showing exactly where pixels differ. Used daily for design QA, client-revision review, photo-edit detection, and A/B visual testing.

If your diff shows changes and you want to know what kindof edit was applied, check the file's metadata with our Metadata Viewer — the EXIF Software field usually names the editor (Photoshop, Lightroom, Snapseed, etc.). For sharpness changes specifically, pair the diff with the AI Image Sharpener to see if one version has been sharpened relative to the other.

Common use cases

What people compare images for

QA a design revision from a client

Client sent back a design mockup with "small tweaks". Drop the original and the revision; diff mode highlights exactly what changed. Great for spotting accidental edits, confirming only the requested areas were touched, and sign-off documentation.

Spot differences between two photos

Two versions of the same scene — before/after retouching, two security- camera frames, product photography with/without reflection removal — the heatmap shows every pixel-level change. Faster than blinking two Photoshop layers on and off.

Verify if a photo has been edited

Got a suspicious image and the original? Diff mode shows where modifications were made — cloned objects leave halo patterns, colour adjustments produce wide faint regions, skin retouching shows concentrated edits around faces. For additional forensic signals, check file metadata with the Metadata Viewer.

Compare compression quality

Testing JPEG quality 80 vs 90 vs 95? WebP vs AVIF at matched file sizes? Drop the output of each compression setting against the original. The score quantifies the quality loss; the heatmap shows where artefacts concentrate (usually sharp edges and areas of fine detail).

Verify AI upscaling or sharpening output

Compare the input and output of any image-processing step to quantify change. Particularly useful for evaluating AI tools — compare the source to the AI Image Upscaler output to see the edge changes; compare sharpened vs unsharpened to see where the sharpener acted.

Design-system visual regression testing

Engineers verifying a CSS change didn't break other parts of the UI drop pre-change and post-change screenshots. Diff mode confirms the only visual changes are where the engineer intended them to be. Free alternative to paid visual-regression SaaS tools for small teams.

Also useful for
compare two imagesspot the difference onlinevisual diff toolbefore after comparisonimage qa toolpixel comparison
Choose your view

When to use slider, side-by-side, or diff

Slider

Subtle differences

Drag-to-reveal overlay. Best when the differences are small and you want to direct your eye to specific spots. Great for client revisions where 'look at the button area' is the workflow.

Side-by-side

Full context at once

Both images visible in parallel at the same time. Best when you need to see the full composition of each without switching back and forth. Ideal for design reviews.

Diff heatmap

Exhaustive change audit

Bright red highlights every changed pixel. Best when you need to find ALL differences — compression QA, edit detection, regression testing. Combine with the numerical score for a summary.

How it works

Four steps, no signup

01
1

Upload two images

Drop the 'before' version and the 'after' version — JPG, PNG, or WEBP up to 10MB each. Different resolutions are fine; the tool normalises to 512×512 for comparison. Nothing is uploaded — everything instant.

02
2

Pick a view mode

Slider reveals one image over the other at your cursor position. Side-by-side shows both at full size in parallel. Diff mode shows a red heatmap highlighting every pixel where the two images differ.

03
3

Read the difference score

A single number from 0 (identical) to 1 (completely different). Under 0.01 usually means 'the same image saved twice'; 0.02–0.1 is typical for compression / format changes; above 0.2 means substantive edits.

04
4

Inspect the heatmap

Diff mode shows exactly where the images differ — text changes glow bright, retouched regions show halos, compression artefacts produce fine speckle. Great for QA on client revisions and design iterations.

Why Texloom

Three view modes, one score, zero uploads

Most online diff tools upload your files and give you one view mode. Ours runs — three view modes and a numerical change score.

Private

Both files stay on your device. No upload, no server, no retention. Works for confidential client work and sensitive internal comparisons.

Numerical + visual

A single 0–1 change score plus a pixel-level heatmap. Summary number for reports, detailed view for investigation.

Any resolution, any aspect

Normalises to 512×512 so you can compare images of wildly different sizes. Practical for comparing a thumb vs a full-res original.

Free, unlimited, commercial use

Compare as many pairs as you need per day. No account, no watermark on the heatmap PNG, no usage restrictions.

FAQ

Frequently asked

Q.What does the difference score mean?
It's the mean absolute pixel difference across both images, normalised 0 to 1. 0.000 = byte-identical. 0.001–0.01 = same image saved twice (JPEG rounding differences). 0.02–0.1 = minor edits, different compression, format change. 0.1–0.3 = noticeable edits, retouching, colour correction. Above 0.3 = substantively different images.
Q.Why are two identical-looking images showing a non-zero score?
Three common causes: (1) JPEG re-encoding — saving a JPG twice produces tiny pixel differences even at max quality; (2) colour-profile changes — same RGB data in sRGB vs Display P3 profiles shows as different; (3) resolution changes — the tool normalises to 512×512 so different source resolutions introduce resampling noise. Scores under 0.01 for 'identical' images are normal.
Q.Can this detect if a photo has been edited?
Yes, if you have an original to compare against. Drop the suspected original and the suspected edit; diff mode shows exactly where changes were made. Can't detect edits without the original reference. For standalone originality analysis, use our EXIF Metadata Viewer to check for Photoshop/Lightroom/Snapseed signatures in the file metadata.
Q.Does size or aspect ratio need to match?
No — the tool resamples both images to 512×512 for comparison. Wildly different aspect ratios (1:1 vs 2:1) will produce a high score just from the different cropping. For accurate change detection, compare images of the same aspect ratio and similar content.
Q.What's the difference between this and Photoshop's 'Difference' layer blend mode?
Similar concept, different delivery. Photoshop's Difference mode gives you a dark image with bright regions wherever pixels differ; it's a full manual workflow. This tool normalises, diffs, and produces a simple score plus a red heatmap in under a second. For quick QA, ours is faster; for pixel-perfect forensic work at full resolution, Photoshop wins.
Q.Is this really free? Are my images uploaded?
100% free, nothing uploaded. Both files stay on your device — the diff runs instantly on your device. No account required for unlimited comparisons. Signup unlocks PDF reports for QA sign-off, Delta E (CIEDE2000) colour-accurate diff, and per-region diff masks (100 credits, no card).
Q.Can I use the diff heatmap commercially?
Yes — it's generated from your files. No watermarks, no restrictions, no attribution required. Download the red heatmap PNG and drop it into a QA report, a design-review doc, or an email to your team.
Q.What file types work best?
JPG, PNG, and WEBP up to 10MB each. PNG preserves exact pixel values (best for UI/design diff), JPG introduces tiny compression differences even when visually identical (still works, just produces small non-zero scores on identical images). Avoid heavily downsampled thumbnails — 512px or larger works best.

Need PDF reports or CIEDE2000 colour metrics? Sign up free — 100 credits, no card.

Get started free
More free tools

Explore the full free suite

Metadata Viewer

Check EXIF + Photoshop signatures.

Try it free

AI Image Sharpener

Fix blurry photos with AI.

Try it free

AI Image Upscaler

Enlarge 2× or 3× with AI.

Try it free
See all free tools

Pro diff & QC tools

Sign up free — 100 credits, 40+ Pro tools, no card.

Create free account

Stay in the loop

Get textile design tips and product updates delivered to your inbox.

Texloom

AI-powered textile design platform. Create seamless patterns, separate colors, and export production-ready files.

Product

  • All Tools
  • Seamless Pattern Maker
  • Color Separation
  • AI Pattern Generator
  • Pantone Matching
  • Textile Printing Software
  • Pricing

Industries

  • Fashion Design
  • Home Textiles
  • Screen Printing
  • Digital Printing
  • Apparel Manufacturing

Resources

  • Free Tools
  • AI Image Upscaler
  • Blog
  • Learn
  • Changelog
  • Roadmap
  • About
  • Editorial Standards
  • FAQ
  • Sitemap

Legal

  • SLA
  • Status
  • Acceptable Use
  • Privacy Policy
  • Terms of Service
  • Refund Policy
  • Cookie Policy
  • Design Security
  • Contact

Compare

  • Texloom vs Photoshop
  • Texloom vs Midjourney
  • Texloom vs NedGraphics
  • Texloom vs PatternedAI
  • Texloom vs Kaledo

© 2026 Texloom Studio. All rights reserved.

Your designs are private — never shared, never used for AI training
SitemapAll systems operational