How to find and safely remove duplicate files on Windows

Duplicate files can waste time and make folders harder to navigate, but two files with the same name are not necessarily the same file—and two identical files are not necessarily safe to remove. Before cleaning up, make a backup, check where each copy is used, and choose a tool that verifies content and gives you a way to restore files.

Start with a backup

A recovery feature inside a duplicate finder is useful, but it is not a backup. A recovery folder on the same drive can be affected by drive failure, accidental formatting, or another problem with that drive. Back up important files to a separate drive or another trusted location first. Windows File History can save versions of personal files to an external drive or network location; see Microsoft's backup and restore with File History guide.

Compare contents, not just filenames

Names and file sizes are useful clues, not proof. Files called report-final.docx and report-final-copy.docx might be identical, while two files named photo.jpg could contain different pictures. A duplicate checker should compare the actual content before it labels files as duplicates.

Hannun first groups candidates with the same size. For larger files it samples a few portions to rule out obvious differences, then reads and compares each candidate's full SHA-256 hash before showing a verified duplicate group. Sampling can reject files that differ; it does not by itself certify a match. Before cleanup, Hannun also requires a completed comparison and checks the selected files again during the move.

Limit the scan to folders you understand

Choose a specific folder or collection with a clear purpose, such as an old photo import or a Downloads folder. Do not start by scanning system folders, application directories, development projects, or a whole drive you do not understand. Identical files can be required in more than one location: software projects, game installations, backup sets, and synced folders may rely on separate copies.

Check that the scan covers the folders you intended and has finished. A partial scan, inaccessible folder, or skipped file can leave out copies. Hannun reports incomplete scans and skipped paths; cleanup is available only after a completed comparison. The current recovery workflow supports direct local NTFS drives. Network, mapped, substituted, and other file systems are not supported for this move; if Hannun cannot safely move a selected file, that file stays in place.

Review each group before choosing a copy

For every verified group, inspect the full paths and decide which location should remain. Keep the copy in the folder that programs or people expect to use. Consider dates, project context, and whether one copy belongs to a backup or archive. Do not automatically keep the newest file: timestamps can change when files are copied, restored, or synchronized.

If you cannot tell why a copy exists, leave it alone until you can check. The goal is not to reduce a number on a screen; it is to remove only a copy you understand and can recover if your decision turns out to be wrong.

Move selected extras to recovery, then verify

Hannun asks you to select specific copies and confirm the cleanup. It moves each verified file into a .hannun-recovery folder on the same drive and saves recovery records. Open the Recovery view to inspect or restore an individual file. Restoration checks the file and its destination and refuses to overwrite an occupied path.

This is a reversible move, not permanent deletion. The files continue to use disk space while they are in recovery. Keep the recovery folder and Hannun's app data until you no longer need the restore records. If you want to reclaim space later, first verify the kept files and any backups; then decide separately what to remove. Hannun does not automatically erase recovered files or send them to the Windows Recycle Bin during duplicate cleanup.

A careful cleanup checklist

Back up important files to a separate drive or trusted location.
Scan one understood folder or collection, and wait for a completed result.
Check warnings and skipped paths; repeat later if an important location was unavailable.
Review each candidate's path and context, not just its name or timestamp.
Select only copies you can identify, read the confirmation, and move them to recovery.
Open the kept files and check that your normal workflow still works.
Restore anything you miss. Keep recovery data until you are confident.

I'm the maker of Hannun, an offline Windows file finder that lets you choose folders and compare verified duplicate content. The app's installer is unsigned, so Windows may show “Windows protected your PC”; choose More info → Run anyway only if you intend to install it. Product details and download: hannun.pages.dev.