Find duplicate photos online without uploading them

What “online” means for DupShelf

The app is delivered as a website so you can bookmark it and use it without an installer. “Online” refers to how you access the tool, not where your files are processed. Enumeration, hashing, grouping, and preview all run in your browser tab using the File System Access API on supported browsers. Our servers do not receive your image bytes for duplicate detection.

How duplicate detection works

DupShelf reads each supported image file from the folder you choose, computes a SHA-256 hash of the raw file contents, and groups files that share the same hash. That is exact duplicate detection: byte-for-byte identity. Two files match even if one is named vacation.jpg and the other IMG_8842 (1).jpg in a different subfolder. This is the safest first cleanup pass because every grouped file is provably identical to another in the group.

Who this workflow fits

If you need similar-image matching (burst shots, re-compressed memes), DupShelf today focuses on exact copies first. A labeled similar mode is planned later.

• Anyone cleaning Downloads, Pictures, or a backup folder on a laptop
• Users on slow or metered internet who cannot upload terabytes
• People who want to verify privacy in DevTools → Network during a test scan
• Families with WhatsApp forwards and renamed copies in one directory

Step-by-step: first scan

Start with a smaller folder if your machine has limited RAM. Session restore can reload results without re-hashing if the tab closes mid-review.

• Open DupShelf in Chrome or Edge on desktop
• Click Choose folder and select your library root (e.g. Downloads)
• Wait for hashing to finish; use progress and cancel if needed
• Expand each duplicate group and preview thumbnails
• Mark one keeper per group (often largest resolution)
• Move non-keepers to dupshelf-duplicate-images or export CSV
• Delete the duplicate folder in Explorer or Finder only after review

Browser and device limits

Full recursive folder scan and move-to-folder require Chromium desktop browsers (Chrome or Edge). Safari and Firefox can add files manually but do not support the same folder-picker workflow. Mobile browsers are best for small batches; export photos to a PC for a full-library pass.

Privacy checklist before a big library

Before scanning wedding or medical images, open Network once during a test folder scan. You should not see multi-megabyte POST requests of your photos. Try offline mode after the first load—the tool should still open. Read our privacy page for how permissions work. DupShelf never auto-deletes; you confirm every deletion outside the app.

How long an online-local scan takes

A few hundred images may finish in under a minute on an SSD. Five thousand images on a USB hard drive can take twenty to forty minutes because every file is read fully for hashing. Progress and cancel are there so you can stop a mistaken folder pick. After hashing, review is instant—you are not waiting on a server queue.

Troubleshooting when nothing groups

If you expected duplicates but see zero groups, confirm the files are byte-identical (not just similar). Check that both copies are inside the folder you selected. Cloud-only placeholders must be downloaded first. Different image formats or re-compressed WhatsApp forwards will not match in exact mode.

Summary and next steps

Online access with local processing is the sweet spot for trust and convenience. Share this guide with relatives who are nervous about cloud cleaners but comfortable opening Chrome. When they see duplicate groups with their own filenames and paths, skepticism drops. Encourage a test folder of twenty images before a 10k library. Pair this workflow with good backup and you have a complete personal media hygiene stack without subscriptions. Revisit after every major import—phone swap, vacation, wedding season—and the tool stays familiar because no account resets your history.