Duplicate Image Finder

Each image is downscaled to 16x16 grayscale pixels and converted into a perceptual hash of brightness bits. Images with identical hashes are grouped as duplicates.

Up to 50 files can be added at once through the batch drop zone, and all of them are hashed and compared locally.

The perceptual hash tolerates resizing and slight edits, but grouping uses strict hash equality, so heavily edited images may not match.

No. Hashing is computed locally on a canvas element, so the images never leave your browser, even when you compare a whole batch of fifty files.

Duplicate groups are listed with the first image labeled original and the rest labeled duplicates, each with its file name and thumbnail.

A green message says No duplicates found once every image has a unique perceptual hash, and every file is listed with its own thumbnail so you can verify the result.