BYTETOOLS

Duplicate Photo Finder

Find duplicate and near-duplicate photos in a folder. Each image gets a 64-bit perceptual hash, and close matches are grouped with previews.

Drop a folder of photos, or click to select many at onceUp to 300 images — all fingerprinted locally, nothing is uploaded

Each photo is reduced to a 64-bit difference hash, so matching is about structure rather than exact bytes: a resized, re-compressed or lightly cropped copy still lands within a few bits of the original. A threshold of 0 finds near-identical files, 5 is a good default, and above about 12 unrelated photos start appearing together.

What is the Duplicate Photo Finder?

Duplicate photos are rarely identical files. A copy that has been resized, re-saved at a different JPEG quality or exported by a different app has completely different bytes, so checksums miss it.

  • 64-bit difference hash per image — matches resized and re-compressed copies, not just identical files
  • Adjustable Hamming-distance threshold from 0 to 20 bits
  • Groups shown with thumbnails, dimensions, file size and each file's hash
  • Totals for duplicate groups, extra copies and the disk space those copies take
  • Copyable text report of every group
  • Runs entirely offline in your browser — no photos are uploaded

How to use the Duplicate Photo Finder

  1. 1

    Drop in a folder of photos, or click to select many files at once.

  2. 2

    Leave the match threshold at 5 to start — that catches resized and re-compressed copies.

  3. 3

    Read the duplicate groups: each card shows the file name, dimensions, size, hash and how many bits it differs from the first image in its group.

  4. 4

    Raise the threshold to catch looser matches such as crops, or lower it towards 0 for near-identical files only.

  5. 5

    Copy the text report if you want a list to work through in your file manager.

About the Duplicate Photo Finder

Duplicate photos are rarely identical files. A copy that has been resized, re-saved at a different JPEG quality or exported by a different app has completely different bytes, so checksums miss it. Perceptual hashing solves that by describing what the image looks like rather than what it contains.

Every photo you drop in is reduced to greyscale and turned into a 64-bit difference hash, where each bit records whether one pixel is brighter than the one to its right. Two images are grouped together when their hashes differ by fewer bits than your threshold, so resized and recompressed copies still land in the same group as the original.

Groups are shown side by side with each file's dimensions, size and hash so you can decide which copy to keep, and there is a copyable text report of the whole result. Nothing is uploaded — the images are decoded and hashed entirely in your browser.

Frequently asked questions

How does a duplicate photo finder detect resized copies?

It compares perceptual hashes rather than file bytes. Each image is shrunk to a tiny greyscale grid and turned into 64 bits describing its brightness pattern, and resizing or re-compressing a photo barely changes that pattern, so the hashes stay within a few bits of each other.

What threshold should I use?

Zero finds only images with an identical hash, 5 is a good general default that catches resizes and quality changes, and 10 or more starts matching crops and edits. Above about 12 unrelated photos with similar composition begin appearing together.

Can it delete the duplicates it finds?

No. A web page cannot delete files from your disk, and that is deliberate — nothing here can modify or remove anything. Use the previews and the copyable report to decide which copies to keep, then delete them yourself in your file manager.

Are my photos uploaded anywhere?

No. Every image is decoded and hashed by your own browser, and no data of any kind is sent over the network. You can turn off your connection before you start and the tool works exactly the same.

Why do two clearly different photos appear in the same group?

A 64-bit hash is a summary, so two photos with a similar overall brightness layout — two sunsets, two documents on white paper — can land close together. Lower the threshold to tighten the matching, and always check the thumbnails before deleting anything.

Related tools