Windows 11 does not have a duplicate file finder. Storage Sense clears temp files, old updates and Recycle Bin contents — it will not touch the three copies of your phone backup sitting in Downloads, Desktop and D:\Backup.

So people go looking for a tool, and the first thing they find is a wall of "1-click, 50 GB free!" apps that pre-select every duplicate for deletion. That pre-selection is the part that costs you a weekend.

Here's the version that doesn't.

Why "same name" and "same size" both fail

There are three ways to decide two files are the same, and only one of them is safe.

Name matching finds report (1).docx next to report.docx. It misses IMG_4821.jpg and 2024-trip/IMG_4821.jpg, which are the same photo, and it flags 2024-Q1.xlsx against 2024-Q2.xlsx if someone named them carelessly. Wrong in both directions.

Size matching is better but still wrong. Two videos can both be exactly 512 MB and have nothing in common. Size narrows the candidates; it does not confirm anything.

Content hashing is the only check that actually answers the question. The tool reads the bytes and computes a hash. If two files produce the same hash, they are identical — not similar, identical.

Good tools do this in three passes, cheapest first: group by size, hash the first 4 KB of everything that survives, and only then read the full file for the handful that still match. That is why a scan over a few hundred thousand files finishes in minutes instead of hours, and why the result isn't a guess.

The four steps

1. Point it at folders, not the whole drive

Scanning C:\ sounds thorough and is a bad first run. It pulls in Windows\WinSxS, Program Files and every app cache on the machine, and you get thousands of "duplicates" that are supposed to be there.

Start with the three folders duplicates actually accumulate in: Downloads, Pictures, Desktop. Add a backup drive if you keep one.

2. Scan, then read the groups — don't delete anything yet

Each line is one group: files with identical content in different places. The size next to a group is what you get back if you keep one copy of each.

3. Open a group and pick the keeper yourself

This is where most cleaners decide for you. Don't let them. Open each group and check three things:

  • Which path is the original? The file in Pictures\2024\ is usually the real one; the one in Downloads\ is usually the copy.
  • Does anything reference it? A photo referenced by a Lightroom catalog, or an attachment embedded in a project file, breaks if you delete it — even though the copy looks identical.
  • Is it really a duplicate? Same content for a different purpose is rare but real: two identical blank templates, for instance.

If a group is ambiguous, skip it. There is no prize for cleaning everything in one pass.

4. Delete to the Recycle Bin, not permanently

The confirmation dialog should tell you where the files go. If it says "permanently delete", close it. Recycle Bin means a mistake costs you a right-click instead of a file.

Empty the Recycle Bin a week later, after you have used the machine normally. That week is the entire safety margin, and it is worth the disk space.

What this won't fix

Being straight about the limits, since the "50 GB in one click" apps aren't:

  • It won't find near-duplicates. A photo resized to 50% is a different file with a different hash. So is a video re-encoded at a lower bitrate. Content hashing is exact-match only.
  • It won't clean up Windows itself. WinSxS, System Volume Information and hibernation files are a different problem with a different tool (Disk Cleanup, or Dism /Online /Cleanup-Image).
  • It won't tell you what's safe to delete. That judgement is step 3, and no tool can take it over.

One more thing worth knowing: files that look like duplicates but are actually hard links — the same data with two names — free zero bytes when deleted. A tool that counts them as duplicates is lying to you about how much space you will get back. Good ones detect and skip them.

The short version

Windows will not find duplicates for you. Sort by size to narrow it down, hash to confirm, keep the copy in the folder you would look in first, and send the rest to the Recycle Bin instead of deleting them.

I use FileDedupe for this. It is free, needs no account, and deletes to the Recycle Bin by default. The Pro version adds CSV export, saved exclusion rules and command-line use — you do not need any of that for a one-off cleanup.