Local duplicate finder comparison: DupShelf vs Scanly vs PixDuplicate

What all three have in common

DupShelf, Scanly’s duplicate scanner, and PixDuplicate all advertise client-side processing: the marketing promise is that your photos are analyzed in the browser tab, not on a remote server. That is a meaningful privacy improvement over classic upload-based cleaners. It is not a guarantee by itself—you should still open DevTools → Network during a scan and confirm that image bytes are not POSTed to a third-party domain. All three also work best on desktop Chrome or Edge for folder-scale workflows, with more limited paths on Safari and mobile.

Feature comparison at a glance

Use this table as a decision aid, not a scorecard. “Better” depends on whether you need exact copies or similar shots, and how many files you have.

• Algorithm — DupShelf: SHA-256 (exact bytes). Scanly duplicate scanner: dHash (perceptual). PixDuplicate: perceptual hashing with quick/deep modes.
• Typical batch size — DupShelf: thousands of images in one folder scan. Scanly: on the order of ~50 images per duplicate-scanner session. PixDuplicate: varies by mode; oriented toward smaller sets unless you work in batches.
• Folder scan — DupShelf: full recursive folder in Chrome/Edge. Scanly/PixDuplicate: strong for picked files; folder workflows depend on browser support.
• Match type — DupShelf: renamed copies, re-exports, same bytes different extension. Perceptual tools: resized crops, re-compressed JPEGs, near-matches (more false positives).
• Cleanup action — DupShelf: move to dupshelf-duplicate-images or CSV export; never auto-deletes. PixDuplicate: move or permanent delete options. Choose based on risk tolerance.
• Extras — Scanly is one tool inside a large scanner hub (EXIF, QR, OCR). PixDuplicate offers a browser extension. DupShelf focuses only on duplicate photos at folder scale.

Exact duplicates vs similar photos (read this first)

Most bad duplicate-cleaner experiences come from mixing these two goals. Exact duplicate detection asks: are the file bytes identical? If yes, keeping one copy is logically safe. Similar-image detection asks: do these look alike to an algorithm? That can group burst shots, crops, and edits you may still want. DupShelf ships exact mode only today on purpose—it is the conservative first pass. Scanly and PixDuplicate lean perceptual for their duplicate features, which helps when files are not byte-identical but hurts when you only wanted true copies.

When DupShelf is the better fit

Choose DupShelf when your pain is storage wasted on true copies—WhatsApp forwards saved under new names, backup folders copied twice, sync conflicts that left “Copy of…” files next to originals, or export folders from Google Photos with redundant downloads.

• You need to scan all of Downloads or Pictures, not 50 files at a time
• You want provable matches before moving anything (byte-for-byte)
• You prefer moving extras to a review folder or exporting CSV over one-click delete
• You are cleaning a library on disk or an external drive in Chrome/Edge
• You want session restore after a long scan (results saved in the browser)

When Scanly or PixDuplicate may fit better

Perceptual tools earn their place when duplicates are not identical files. Example: you exported the same photo as a high-res JPG and a smaller WhatsApp-compressed JPG—bytes differ, story is the same. Or you want to collapse burst shots that are visually almost identical. Trade-offs: stricter batch limits on free tiers, more manual review for false positives, and you should read whether delete is permanent.

• Small album (dozens of images) where upload-style batch limits are fine
• You need a similarity slider (strict vs loose) more than cryptographic certainty
• You want near-duplicate detection, not just renamed copies
• You already use Scanly for EXIF/OCR and want duplicate scan in the same ecosystem

Privacy: how to verify any “local” tool

Before scanning wedding albums, medical images, or work documents, run a three-minute audit. Load the tool once, open Network, start a scan on a test folder, and filter by your domain. You should see page assets and maybe analytics—not multi-megabyte POST bodies of your images. Try working offline after the first load; a genuinely client-side tool should still open. Read the privacy policy for words like upload, retention, and server-side processing. DupShelf documents folder read/write permissions separately because move-to-folder requires explicit write consent in Chromium browsers.

Recommended two-pass workflow

Power users often combine approaches instead of betting on one app. Pass one: run an exact dedup tool on the whole folder (DupShelf) and move provable duplicates to a quarantine folder. Pass two: if you still need space or still see visual clutter, run a perceptual tool on a smaller subset or on remaining folders. That order minimizes accidental deletion of unique photos while still catching re-compressed copies the first pass missed.

Try DupShelf on your folder

If your library sounds like the DupShelf use cases above, open the workbench, choose a folder in Chrome or Edge, and review groups before you delete anything in Finder or Explorer. Free, no account, no upload.