Similar photos vs duplicates on iPhone: what is safe to merge?
Use Apple's Duplicates collection for sets that Photos identifies and offers to merge. Treat visually similar shots and burst frames differently: compare them, because expression, focus, framing, edit history, or meaning can vary even when two thumbnails look almost the same.
Merge Apple-identified duplicates; review similar moments manually.
Apple says Photos combines the highest image quality and relevant data when you use Merge on an identified duplicate set, then moves the other copies to Recently Deleted. That workflow is not the same as choosing between two genuinely different photos. A Swoof similarity group may contain separate captures, so Swoof does not apply Apple's merge semantics or pick an automatic winner.
Three groups that look alike but require different decisions.
Group
What it can contain
Safer action
Apple Duplicates
A set Photos has identified and presents with a Merge control.
Inspect the set and use Apple's Merge workflow if it matches your intent.
Visually similar photos
Different moments, crops, exposures, expressions, or edits.
Compare at full size and choose manually; keep both when each has value.
Burst series
Rapid captures from one sequence, often with small timing differences.
Check faces, motion, focus, and the peak moment before removing frames.
How to use Apple's Duplicates collection.
Open Photos and tap Collections.
Scroll to Utilities, then open Duplicates.
Inspect the images and videos in the proposed set.
Tap Merge only when Apple's proposed set matches what you intend to consolidate.
Apple notes that the Duplicates collection does not appear when Photos has no duplicate set to show, and detection can take time. Do not interpret a missing collection as proof that every image is unique. It only means Photos is not currently presenting a mergeable set there.
Swoof side by side
Move the split and inspect the same part of each image.
The comparison slider keeps both candidates on one surface. Move the divider to check the intended subject, switch challengers within the group, keep both when that is honest, or mark a confirmed photo for Swoof Trash. Swoof does not call either image the winner.
What to compare before choosing.
Faces: eyes, expression, and whether anyone is obscured.
Focus: the intended subject, not automatically the whole frame.
Motion: gesture, body position, and whether blur adds or removes meaning.
Framing: cropped hands, horizons, background distractions, and usable space.
Edit and format: an edited version may intentionally differ from its source.
Context: a technically weaker image may be the only record of a moment.
Similarity is useful for discovery, not for deciding personal value. A strict detector can miss related shots; a broad detector can group items that you would not consider interchangeable. Swoof exposes similarity strictness but makes no promise to find every possible relationship.
Swoof Trash keeps comparison separate from deletion.
When you choose to remove a candidate, Swoof queues it first. You can undo within the comparison flow or remove it from Swoof Trash before the final action. Empty Trash is the distinct step that requests deletion through Apple Photos.
Apple normally moves deleted items to Recently Deleted for up to 30 days unless they are permanently removed sooner. If iCloud Photos is enabled, deletion can synchronize across other devices. Review the queue and your backup state instead of treating the recovery window as a guarantee.
No. A set that Apple identifies in its Duplicates collection is eligible for Apple's Merge workflow. A Swoof similarity group can contain different moments, crops, edits, or burst frames and needs a manual keep-or-trash decision.
Does Swoof merge exact duplicates?
No. Use Apple's Duplicates collection for Apple's merge behavior. Swoof provides a separate side-by-side review for visually similar photos and does not claim to reproduce Apple's merge.
Does Swoof automatically choose the best shot?
No. Sharpness and visual similarity can help organize candidates, but you decide which expression, framing, edit, or moment matters.
Will Swoof find every similar or duplicate-looking image?
No completeness guarantee is made. Results depend on available on-device representations, burst information, and the selected similarity strictness.
Ready when you are
Use automation to find the group. Keep the judgment human.
Open Swoof and review one small, manageable part of your library.
