How does the image duplicate finder work?

Each image is decoded and reduced to a 64-bit perceptual fingerprint (a difference hash) plus a SHA-256 of the file bytes. Images with near-identical fingerprints are grouped, so it catches resized, recompressed, renamed, and lightly edited copies - not just byte-identical files.

Is my image uploaded to a server?

No. Hashing and comparison run entirely in your browser using the Canvas and Web Crypto APIs. Your images are never uploaded, stored, or sent anywhere - close the tab and nothing remains.

What is the difference between exact and near-duplicates?

Exact duplicates are byte-identical (same SHA-256) or pixel-identical files. Near-duplicates look the same to a person but differ in size, compression, format, or minor edits. Use the sensitivity setting to control how aggressively similar images are grouped.

Which image is recommended to keep?

Within each duplicate group the tool recommends keeping the highest-resolution file, then the largest by bytes, so you preserve the best master. You can override the choice, then download the kept set as a ZIP.

How many images can I check at once?

It is built for catalog-scale batches - drop a whole folder. Everything runs locally, so very large sets are limited mainly by your device's memory rather than any upload limit.

Catalog hygiene

Find duplicate and near-duplicate images.

Drop a folder of photos and instantly see exact and visually similar duplicates grouped together, with one click to keep the best copy - all in your browser.

Open duplicate finder Browse all tools

Image Duplicate Finder - Find Duplicate Photos - CatalogPrep

Find exact and near-duplicate images in a batch - resized, recompressed or edited copies - using perceptual hashing. Free, in your browser, no upload.

Frequently asked questions

How does the image duplicate finder work?: Each image is decoded and reduced to a 64-bit perceptual fingerprint (a difference hash) plus a SHA-256 of the file bytes. Images with near-identical fingerprints are grouped, so it catches resized, recompressed, renamed, and lightly edited copies - not just byte-identical files.
Is my image uploaded to a server?: No. Hashing and comparison run entirely in your browser using the Canvas and Web Crypto APIs. Your images are never uploaded, stored, or sent anywhere - close the tab and nothing remains.
What is the difference between exact and near-duplicates?: Exact duplicates are byte-identical (same SHA-256) or pixel-identical files. Near-duplicates look the same to a person but differ in size, compression, format, or minor edits. Use the sensitivity setting to control how aggressively similar images are grouped.
Which image is recommended to keep?: Within each duplicate group the tool recommends keeping the highest-resolution file, then the largest by bytes, so you preserve the best master. You can override the choice, then download the kept set as a ZIP.
How many images can I check at once?: It is built for catalog-scale batches - drop a whole folder. Everything runs locally, so very large sets are limited mainly by your device's memory rather than any upload limit.