🔁Duplicate Image Finder

Supplier folders arrive with the same photo saved three times at different sizes, phone exports duplicate every shot you edited, and a year of listing work leaves thousands of near-identical files. Uploading duplicates wastes listing slots and storage; deleting by eye takes hours. The Duplicate Image Finder is a similar image finder that groups exact and near-duplicate photos in a batch, lets you keep the best of each group and hands back the cleaned set, so you can remove duplicate images with confidence and find duplicate photos you did not know you had.

How it works

1. Drop the folder, up to 200 files at a time.
2. Each image is downsampled to a tiny greyscale grid and turned into a 64-bit perceptual hash (dHash) that describes the brightness gradient pattern. Resizing, re-compression, format changes and small crops barely change the hash; different photos change it a lot.
3. Hashes are compared in pairs and images closer than the similarity threshold are grouped. The default of about 90% catches resized and re-saved copies without merging different colour variants.
4. Each group shows its members with dimensions and file size. Use Keep best to select the largest and sharpest automatically, or pick by hand.
5. Download a ZIP of the kept files only, or a CSV of the groups for your records.

Because the hash is 64 bits, comparing a few hundred images takes milliseconds; the slow part is decoding, which happens in a worker.

Why it matters

Duplicates cost more than disk space:

• Etsy allows 10 photos per listing and Daraz 8; a duplicate in that set is a wasted slot that could have shown another angle.
• Shopify accepts 250 images per product, and many themes show all of them, so a repeated image is visible to customers.
• Uploading the small copy instead of the large one is a common mistake in bulk uploads; the finder highlights the largest file in every group so the Amazon 1000 px zoom minimum is met by the right version.
• Repeated main images across separate listings can trigger duplicate-listing reviews.

Tips

• Lower the threshold to 80 to 85% to catch aggressive crops and heavy edits; raise it to 95% or more when your catalogue has legitimate variants that differ only in colour.
• Keep best prefers pixel count first and file size second; check the choice when a large file is a badly compressed upscale.
• Run the finder before renaming or converting so file names still indicate the source.
• After cleaning, sort the survivors into folders with the Image Sorter & Organiser and rename them with the Bulk Rename + Convert tool.
• When two images are in the same group and you cannot tell them apart, open them in the Image Compare / Diff tool for a pixel-level answer.
• Confirm the kept files meet each marketplace's minimum with the Image Dimension & Size Checker.

Privacy

Hashing and grouping run entirely in your browser. Your images never leave your browser, and the hashes are discarded when the tab closes.