Duplicate File Finder

About the duplicate file finder

Duplicates pile up quietly. You copy a photo folder "just in case", a download runs twice and leaves invoice (1).pdf, a backup gets restored into the wrong place, and years later a drive is full of files you already have. This tool finds them: pick a folder, and it reports every set of files that are identical, how many copies exist, and how much space you would get back.

Identical means byte for byte, not "looks similar". Two photos of the same moment are two different files and are never reported as duplicates. Two copies of the same file are, no matter what they are called or which subfolder they sit in.

The folder is read inside your browser. Nothing is uploaded, and the tool has read-only access — it cannot delete anything, which is exactly why it hands you a list to check rather than quietly removing files.

How it decides two files are the same

Comparing every file against every other file would take forever on a big folder, so it works in three passes and throws work away at each one.

• Size first. Two files of different lengths cannot possibly be identical. This alone rules out the overwhelming majority of a typical folder and costs nothing — the sizes are already known
• Then the first 64 KB. Files that are the same size but different usually differ near the start — camera JPEGs, padded archives, documents from a template. Hashing just the head separates them without reading the rest
• Then the whole file. Only for the handful that still match, so the expensive read happens on real candidates and nothing else

The hash is SHA-256, computed by your browser's built-in cryptography. For two different files to be reported as duplicates they would have to share a 256-bit digest, which does not happen in practice.

How to use it

• Click Choose a folder and pick the one to check. Your browser will ask permission to read it
• Wait for the passes to finish — the progress bar covers the hashing, which is the slow part
• Read the groups. Each one shows the copies it found, with the one to keep marked
• Choose which copy to keep — newest, oldest, shortest path, or closest to the top folder
• Take the result — copy the list, download a CSV, or download a delete script

Choosing which copy to keep

The tool never decides for you which file matters; it just marks one per group as the keeper so the rest are labelled clearly.

• Newest — keep the most recently modified copy. The usual choice when you have been working in one place
• Oldest — keep the original. Useful when copies were made by software that reset the timestamps
• Shortest path — keep the one with the least deeply nested path, usually the "real" home rather than a backup folder
• Closest to the top folder — keep whatever sits nearest the folder you picked

Changing this only re-labels the groups. It does not re-scan, so you can flip between them instantly.

Narrowing the search

• Ignore files under (KB) — skip small files. A folder of source code or thumbnails will have hundreds of tiny identical files that are not worth the noise; set 100 to look only at things worth reclaiming
• Only these types — type jpg, png for photos, mp4, mov for video, pdf for documents. Dots optional, case ignored

Zero-byte files are always skipped. Every empty file is byte-identical to every other empty file, so reporting them would bury the real results.

The delete script

Deleting is the one genuinely destructive thing here, so the script arrives with every delete commented out. You open it, read the list, remove the REM (Windows) or # (Mac and Linux) from the lines you actually want, and run it yourself. Nothing happens until you do.

That is deliberate. A web page that could delete files on your PC the moment you clicked a button would be a bad idea no matter how careful the code was — so this one cannot, and the script makes the last step yours.

What people use it for

• Reclaiming space on a photo drive that has been copied between computers over the years
• Cleaning a Downloads folder full of file (1).pdf, file (2).pdf
• Finding music duplicated across several imports of the same library
• Checking whether a backup folder is holding a second copy of things you never deleted
• Working out what is actually taking up space before buying a bigger drive

Two ways your browser reads a folder

Chrome, Edge and Opera support a proper folder picker. The tool walks the folder itself and reads each file's bytes directly through the handle, which is the fastest route and the only one that sees empty subfolders.

Firefox and Safari use the older folder upload control. The browser hands over every file in the tree at once. Duplicate detection works exactly the same — every file is still read and hashed in the page — but the whole tree is loaded up front, so a very large folder takes longer to get started.

Good to know

• Nothing is uploaded. Reading and hashing both happen in the page — you can watch the network tab
• The tool has read-only access. It cannot create, rename, move or delete a single file
• Hashing reads every candidate file, so a folder with many large same-size files takes a while. The size and head-hash passes exist precisely to keep that rare
• Very large folders are capped at 200,000 items so the page cannot freeze
• Files your browser refuses to open (locked, in use, or in a protected system folder) are skipped rather than reported wrongly
• The CSV includes a byte-order mark so accented and non-Latin filenames open correctly in Excel

Related tools

• Folder & File List — list everything in a folder and copy it
• Compare Two Folders — see what differs between two folders
• Bulk File Rename — rename a whole folder at once
• Hash Generator — checksum a single piece of text or a file