Duplicate files are the easiest storage win on a Windows PC. You probably have three copies of that holiday photo album, the installer for every app you've ever downloaded, and a dozen "Untitled (1)", "Untitled (2)" screenshots in Downloads. This guide shows you how to find them and safely delete the extras — for free, without risking system files.
Quick warning before we start. "Duplicate finder" is exactly the kind of utility that goes wrong if it's too aggressive — Windows uses hard links and shared system files in ways that look like duplicates but aren't. The safe approach is to scan the folders you own (Pictures, Downloads, Documents, etc.) and stay out of the system folders.
Why filename matching isn't enough
Plenty of free tools match by filename only. That's almost useless — IMG_2034.jpg might be two totally different photos if your camera resets counters, and two copies of the same song may have wildly different filenames.
The reliable approach is hash-based comparison: the file's contents are hashed (usually SHA-256 or xxHash) and two files are duplicates only if their hashes match. That way you find true byte-for-byte duplicates, not similarly-named files.
What to scan
- Downloads folder — jackpot. Installers, ZIP files, PDFs.
- Pictures / Photos library — backups of backups of backups.
- Videos — can reclaim tens of GB.
- Music — especially if you've moved collections between apps.
- Documents — report drafts, invoice copies, version chaos.
What NOT to scan
C:\Windows— Windows hard-links a lot of files intentionally.Program Files/Program Files (x86)— app DLLs and assets, same story.ProgramData— shared application data.- Drive roots (
C:\) — you'll pull in everything above by accident. - WinSxS, DriverStore, CatRoot — deleting duplicates here breaks Windows Update.
The built-in option (PowerShell, advanced)
If you prefer command-line: Get-ChildItem -Recurse | Get-FileHash | Group-Object Hash | Where-Object Count -gt 1. It works but you then have to review each group manually and delete by hand. Fine for ten files — exhausting for ten thousand.
The fast, safe way: RBS PC Cleaner Duplicate Finder
RBS PC Cleaner has a dedicated Duplicate Finder page:
- Hash-based (true byte-for-byte), with optional same-extension or same-filename filters.
- Always keeps the oldest file and marks newer copies for deletion — so your original stays, newer re-copies go.
- Auto-skips Windows, Program Files, ProgramData, drive roots, WinSxS, DriverStore, CatRoot.
- Confirms before any delete.
- Shows you groups with previews so you can spot-check before committing.
- Moves files to Recycle Bin by default (undo possible) — hard delete is opt-in.
Free. No subscription. No account.
Reclaim expectations
A typical 500 GB-used drive scanned by Duplicate Finder reclaims anywhere from 2 GB to 40 GB depending on how much you've downloaded. Photo libraries migrated across phones are often the single biggest source — every time you change phones, the new device wants to "back up" your old photos and you end up with two or three copies of every shot.
Photo libraries — what to watch out for
Photos are where most people see the biggest gains, but also where it's easiest to delete something you wanted. A few specific things:
- EXIF-stripped copies are NOT duplicates of the original. If you shared a photo via WhatsApp or Instagram and saved the result, the file size and pixel data look identical but the metadata (camera info, GPS) is gone. Hash-based comparison correctly identifies these as different files. Some "smart" duplicate tools incorrectly flag them as the same — be careful which one you use.
- Edited copies look different to a hash, even if visually similar. A cropped or filtered photo will have a completely different hash from the original. Hash-based tools won't see them as duplicates. That's the right behaviour but it means hash dedup won't catch "the same photo, slightly different edits".
- Live Photos and HEIC have multiple components. An iPhone Live Photo is a HEIC + a small MOV. If you copy the album twice, both files duplicate. The video components are easy to miss because they don't show in Photos preview.
"What about visually similar but not identical photos?"
That's a different problem called perceptual hashing — finding photos that look the same to a human even if the pixels differ slightly. It's useful for cleaning up bursts ("I took 8 nearly-identical photos of the same sunset, keep the best") but it's also dangerous, because deciding what counts as "similar enough" is subjective. RBS PC Cleaner deliberately doesn't do perceptual hashing — for that I'd recommend a dedicated photo manager like digiKam, which has it built in and a careful UI for reviewing matches before deleting.
"Could two different files have the same hash?"
The technical answer: yes, in theory. SHA-256 has 2^256 possible outputs, which is more than the number of atoms in the observable universe. The probability of two random files producing the same hash is about 1 in 10^77. To put that in context, you're more likely to win the lottery while being struck by lightning during a meteor strike. Don't worry about it.
Related articles
- Free Disk Space Analyzer for Windows — find the folders with the most duplicates first.
- How to Clean Windows 11 Without Breaking It — safe cleanup across the whole system.
- Uninstall Multiple Programs at Once on Windows — another big way to reclaim disk space.
- RBS PC Cleaner v1.0.0 Launch — all features in one place.