ImageSlim

Duplicate photos, found without uploading a thing.

ImageSlim finds and removes duplicate photos across OneDrive, Google Drive and Dropbox — and it runs entirely in your browser. There is no backend, and no server ever holds your cloud tokens. Anything you delete goes to your provider's own recycle bin, recoverable for around 30 days.

Exact duplicates are found for free using each provider's own content hash — OneDrive's quickXorHash, Google Drive's md5Checksum — so nothing has to be downloaded just to compare it. Visually similar photos are the harder case: resized, recompressed, renamed, or the same shot sitting in two different clouds. Those are caught with a perceptual hash and Hamming-distance clustering, computed off the main thread in a web worker and cached in IndexedDB so a second scan is fast.

Once it has the groups, a smart-keep heuristic suggests which copy to keep — highest resolution, then largest, then oldest, then shortest filename — and nothing is deleted until you confirm. You can dial match sensitivity between strict and loose, and scan everything or a faster quarter sample.

• Connect OneDrive, Google Drive, or Dropbox — no account with us needed
• Exact duplicates via each provider's content hash, with nothing downloaded
• Visually similar photos found with perceptual hashing and clustering
• Catches the same photo stored across two different clouds
• Smart-keep suggests which copy to keep — you confirm before anything goes
• Deletions land in your provider's recycle bin, recoverable ~30 days
• Match sensitivity: strict, balanced, or loose
• Scan everything, or a half or quarter sample for speed
• Computed hashes cached in IndexedDB, so re-scans are fast
• Runs entirely client-side — no backend ever sees your tokens or your photos
• Google Photos and iCloud Photos are not supported — those APIs do not allow it