Find duplicate files, verified byte-for-byte

Name and size first, then every byte

Files are first grouped by name and size — a fast pass to find candidates. Turn on Deep Scan and every candidate is confirmed with a full SHA-256 hash before it counts as a duplicate. Symlinks pointing at the same target are detected too, kept in their own group.

Each group gets a confidence score from five signals — folder similarity, naming, timestamps, path proximity, and copy count — so you can tell accidental clutter from something that might be intentional at a glance.

A Keep rule (Manual, Oldest, Newest, or Largest) auto-picks which copy to keep, flagged ★ KEEPER in the group. Mark any single copy Ignore to exclude it from a batch clean without deleting it. And the last remaining copy in any group is always safety-locked — DupSweep will never let a group go to zero.

When this helps

What it does not do

• It does not catch files that only look the same — a re-saved or re-compressed copy has different bytes. For photos, see Photos mode, which compares appearance instead.
• It does not upload your files anywhere. Hashing runs entirely on your computer.
• "Largest" mostly ties, since Files-mode duplicates are byte-identical by construction — it matters more as a tiebreaker after Oldest/Newest.
• Cloud-only placeholders (iCloud / OneDrive files not downloaded) are unreliable. Scan local copies.