Remove duplicate files on Mac without accidentally deleting the wrong copy

Step-by-step duplicate removal

1. Start with Downloads and Desktop. These two folders generate more duplicates than anywhere else — installer .dmg files downloaded twice, documents saved from email alongside earlier versions, archives sitting next to their extracted contents. Scan them first for fast wins before moving to Documents or Photos.
2. Use hash-based comparison, not filename matching. A file renamed from "report_v2.pdf" to "report_final_FINAL.pdf" has the same content — a name-only search would miss it entirely. Tools that compute a SHA-256 hash of each file's actual bytes will find true copies regardless of what they're called or where they live.
3. Review each duplicate group before deleting. For each group of identical files, look at location and modification date side by side. The copy inside a named project folder is usually the keeper; the stray copy buried in Downloads or on the Desktop is the candidate for removal.
4. Handle photo duplicates separately. Never browse inside your .photoslibrary package and delete files manually — the Photos database will lose track of them without freeing space correctly. Use Photos → Albums → Duplicates (added in macOS Ventura) for image deduplication within your library.
5. Delete in batches of 20–50 files at a time. Moving hundreds of files to Trash at once makes it harder to spot mistakes. Smaller batches let you pause, glance at what was just moved, and verify before continuing.
6. Check recovered space after emptying Trash. Open System Settings → General → Storage and confirm the numbers shifted as expected. If they didn't move much, some of the files may have been evicted to iCloud already and weren't counted as local storage anyway.

Not all "duplicates" actually waste space

Understanding what's actually costing you storage helps you focus on the right targets and avoid wasted effort.

• True duplicate copies — two separate files with identical content in different locations. Each takes up its own space. These are the primary cleanup target.
• APFS hardlinks — multiple directory names pointing to the same underlying data block on the APFS filesystem. Time Machine relies on these for efficient incremental backups. They don't consume extra space and should not be removed.
• Near-duplicates — photos or documents that look the same but differ in metadata, edit history, or timestamps. Byte-for-byte they are different files. Only remove these if you've confirmed visually that both versions are truly interchangeable.
• HEIC and JPEG pairs — some older workflows export both formats from the same original. If cross-platform compatibility isn't a concern, keep the HEIC version (roughly half the file size at equal or better quality) and remove the JPEG.
• Archive and extracted folder pairs — the .zip and the folder it produced. Once the extraction is confirmed complete and you no longer need to re-share the archive, the original .zip is safe to remove.

Categories to leave alone during deduplication

Deduplication is most powerful when it stays focused on your personal files. These areas should be skipped entirely:

• Anything inside /System, /Library, or /Applications — these are managed by macOS and app installers.
• Files inside ~/Library/Application Support — apps store essential preferences and state data here that can look like duplicates but isn't.
• The Photos Library.photoslibrary package — never browse or delete inside it using Finder.
• Xcode derived data and project caches — build artifacts can look identical but belong to different build configurations.
• Virtual machine disk images — these can have snapshot-based structures where deleting the "duplicate" snapshot corrupts the entire VM.