DupShelf

DupShelf

Find duplicate images in a folder on your computer

Whether your library lives in Pictures, a client delivery folder, or a messy Downloads tree, duplicate images hide under different names and nested paths. DupShelf lets you point at one root folder and recursively scan every supported image inside it—hashing with SHA-256 and grouping byte-identical files. No upload, no account, and you review every group before anything is moved or deleted.

How folder scan works in the browser

On Chrome or Edge desktop, you grant read access (and optional write access for moves) to a directory handle. The app enumerates files recursively, filters to image types, hashes contents, and builds duplicate groups. Safari and Firefox do not support this full workflow; use manual file add or switch browsers for folder-scale jobs.

Nested subfolders and large libraries

Selecting Pictures or a backup root includes all child folders. That is intentional—duplicates often sit in 2023/ and 2024/ with different names. Progress and cancel help on 10k+ file trees. Low-RAM machines should scan one top-level folder at a time.

Typical workflow end to end

Repeat for another root if you keep photos in multiple locations.

  • Open DupShelf → Choose folder
  • Wait for hashing; expand groups with duplicates
  • Pick keepers (largest file or best path)
  • Move extras to dupshelf-duplicate-images or export CSV
  • Verify in file manager, then delete quarantine folder

Choosing keepers wisely

Default instinct: keep the highest resolution or the file in the “canonical” album folder. For synced drives, keep the copy outside conflict-named files like photo (1).jpg when bytes match the original.

When folder scan is not enough alone

Exact mode will not group re-compressed WhatsApp forwards with the original if bytes differ. Run exact folder scan first, then optional perceptual tooling on remaining clutter.

Catalogs, DAMs, and Lightroom

DupShelf operates on files on disk, not inside Lightroom or Capture One catalogs. Export or sync originals to a folder, dedup, then reimport if needed. Catalog previews may regenerate after dedup—that is normal. Keep catalog backups before bulk file moves.

Team folders and NAS paths

Creative teams with shared NAS should agree one maintainer runs dedup per month on the incoming ingest folder. Conflicting moves from two people at once are avoided by process, not software locks. CSV export helps hand off a duplicate list to a teammate for review.

Mixed photo and video folders

Only images are hashed. Videos in the same tree are skipped. If duplicates are mostly screenshots next to videos, you still get image-only results without errors.

Permissions denied errors

If move fails, the folder may be read-only or on a protected path. Export CSV and delete manually, or rescan with write permission granted in the picker.

Summary and next steps

Folder-level duplicate finding rewards good folder hygiene. If your tree is chaos, scanning still works—but review takes longer because paths are meaningless. Consider light reorganizing before scan: one folder per year, separate WhatsApp exports, and moving obvious trash out of the tree. DupShelf will not judge your structure; it will hash what is there. After cleanup, enforce a simple rule: new imports land in Inbox, monthly scan moves duplicates to quarantine, unique keepers filed into Archive. Nested subfolders are included automatically, so duplicating an entire branch copy-paste style will surface as large groups—exactly what you want before deleting a redundant branch. If you collaborate, agree which machine runs the scan to avoid two people moving the same files. Communication beats conflicting quarantine folders. Pick one root, review groups, move quarantine, verify, delete. Repeat monthly on inbox folders to keep laptops fast.

Questions

Does it scan subfolders?
Yes. Everything under the folder you pick is included.
Can I scan multiple folders in one job?
Pick one root per scan. Run again for another root, or merge copies into one folder first.
What if the folder is on a network drive?
If the OS exposes it to Chrome with read access, it can work. Latency may be high.
Are hidden files included?
Enumeration follows what the browser exposes for the granted directory.
Will symlinks be followed?
Behavior depends on browser and OS. Prefer scanning real folders when unsure.
Can I exclude a subfolder?
Scan a narrower root or move exclusions out before scanning.
Can I scan a ZIP without extracting?
No. Extract archives first; DupShelf reads normal files on disk.

Related guides