Why duplicate finders are usually slow

Finding duplicate files sounds simple — compare every file to every other file — but doing it honestly on a real drive with millions of files means avoiding a combinatorial explosion. Most tools solve this by hashing every single file up front, which means reading the full content of every file on the drive before comparing anything. That's exactly the kind of work a spinning disk, or even a busy SSD, makes painfully slow.

Duplicate Cleaner and similar tools scan the folders you select and compute hashes as part of that scan — thorough, but the scan itself takes time proportional to your data, every single run.
Deskora starts from data it already has: file sizes from the NTFS index. Only files with a size that another file also has become candidates — everything else is eliminated in an instant, with no disk I/O at all.

🧹 Find real duplicates, not false positives

Three stages — size, partial hash, full hash — so only genuine duplicates get flagged.

Download Deskora for free

How the three-stage search works

1️⃣ Group by exact size

Pulled straight from the index Deskora already has — no file is opened, no byte is read. Any file whose size is unique on the whole scope is immediately ruled out; it cannot have a duplicate.

2️⃣ Partial hash on the survivors

For the remaining size-collision candidates, Deskora hashes the first and last 64 KB of each file. Most genuine non-duplicates with a matching size (a coincidence, or files sharing a common header) are eliminated here without reading the whole file.

3️⃣ Full hash for what's left

Only the small number of files that survive both previous stages get a complete, full-content hash — the expensive step, run on the fewest possible files.

🔗 Hardlinks are not false alarms

Two paths pointing at the same physical data on disk are not wasting extra space, and Deskora marks them as hardlinks rather than counting them as duplicates you could "save space" by removing.

How it works, step by step

Open Deskora and switch to the Storage tab, then to Duplicate Finder.
Choose a single drive, a folder, or all indexed volumes.
Stage 1 (size grouping) completes instantly using the existing index.
Stages 2 and 3 run in the background with a visible progress indicator, and can be cancelled at any time.
Results appear as groups of identical files, each showing path, date and size, with the oldest and newest copy clearly marked.
Select the copies you don't need and delete them — always to the Recycle Bin, never permanently — or right-click for the native Explorer context menu.

Deskora compared

Honest comparison. Deskora's duplicate finder lives inside a free desktop organizer rather than as a stand-alone paid product, and it leans on an index that already exists for the size stage specifically.

	Deskora	Duplicate Cleaner
Size grouping without any disk scan	Yes — from the existing NTFS index	No — scans the selected folders first
Three-stage hashing (size → partial → full)	Yes	Configurable hashing options
Hardlinks flagged, not counted as wasted space	Yes	Varies by version
Delete to Recycle Bin only	Yes, always	Configurable
Cancel and resume anytime	Yes	Yes
Price	Free	Free / paid tiers

Who this is for

🎨 Content creators

Exported renders, "final_v2", "final_v2_ACTUAL_final" — creative project folders accumulate exact duplicate assets fast, especially after a few rounds of "just to be safe" copying. A duplicate finder that runs in seconds instead of tying up your machine for half an hour makes it something you'll actually use.

⚙️ Engineers and technical teams

Multiple project copies, backup folders nested inside other backup folders, CAD assemblies exported more than once — poor naming and versioning discipline is the standard problem, and finding the exact duplicates is the fastest first step toward cleaning it up.

📸 Anyone with years of photo backups

Photos copied from a phone, then backed up again, then imported into a photo manager that made its own copy — the same picture can exist in four places without you ever noticing.

Frequently asked questions

Why is the first stage instant?

Because it doesn't touch your disk at all — file sizes are already known from the NTFS index Deskora maintains for search. Grouping by a number you already have costs nothing.

Could two different files with the same size be mistaken for duplicates?

No. Matching size only makes a file a candidate. Stage 2 and stage 3 hash actual file content, so two files that happen to share a size but differ anywhere in their bytes are correctly told apart.

What about zero-byte files?

They're handled as a special case so thousands of empty files don't get treated as one enormous "duplicate group" — that would be technically true and practically useless.

Are hardlinked files deleted as duplicates?

No. A hardlink points at the same physical data as another path, so deleting one copy would not free any space — Deskora marks these clearly instead of listing them as removable duplicates.

Can I stop a duplicate search partway through?

Yes, at any time. Cancelling is immediate and safe — nothing is deleted until you explicitly choose to.

Where do deleted duplicates go?

To the Recycle Bin, always. Nothing is ever deleted permanently by the duplicate finder, and every deletion requires your explicit confirmation — no one-click mass delete.