A safe, batch-based process for cleaning and organizing large file collections —
without needing every file available locally at once.
Built first for a cloud-backed OneDrive folder, but designed to be provider-agnostic.
Work happens in small, resumable batches: duplicates are confirmed by content hash,
only older confirmed copies are moved, and every move step asks for approval first.
Safety rules
Never delete files. Older duplicates are moved, not removed.
Confirm duplicates by content hash, never by filename.
Keep the newest copy where it already lives.
Move older confirmed duplicates into nearby _older_duplicates folders.
Preserve app-managed folders (Attachments, Teams, OneNote, and similar).
Never split CAD, application, license, or package folders.
When in doubt, treat the folder as one indivisible package.