Find Duplicates
Files that are byte-for-byte identical, verified by a full BLAKE3 hash. How matching works
The app, screen by screen
DuoBolt is one app with two engines and a shared way out. This page walks the whole path once, in order, and points at the page that covers each stop properly.
If you have not run a scan yet, start here instead.
1. The dashboard
Section titled “1. The dashboard”
Two engines, and the choice between them is the choice of question:
Similars
Photos that look alike — bursts, exports, re-saves. How grouping works
Recent scans are listed here, and any of them can be run again with the exact setup it used.
Above them sits Continue where you left off, and it only appears when there is something to continue — one card per engine, naming the folders and what the scan found, with Discard and Resume on it. What resuming does is covered under progress.
The rest of the dashboard is counters rather than controls: Lifetime totals every scan you have ever run, and By feature splits the two engines — what is still wasted against what you have already reclaimed on the duplicates side, how many photos are waiting to be reviewed on the other.
2. Scan setup
Section titled “2. Scan setup”Add folders — as many as you like, on any volume — and adjust Scan parameters if the defaults do not fit. The info card underneath summarizes the setup, and mentions only what deviates from the defaults, so a quiet card means a standard scan.
Setting up a duplicates scan Size range, extensions, ignore rules, skip folders, symlinks, performance
Setting up a similars scan The similarity threshold and the group cap
3. Progress
Section titled “3. Progress”Duplicates runs Scan → Prehash → Hash → Finalize; Similars fingerprints and then groups. Both report live counts, and both can be left alone — if the app closes with an open result, the session is saved and offered back on the next launch, selections and all.
4. Results
Section titled “4. Results”Where the work happens. The two screens differ because the questions differ — duplicates is a list ordered by recoverable space, similars is a grid of groups you look at.
Reading the duplicates results Filters, sorts, bulk rules, and the two numbers in the footer
Reading the Similars results Tightness, Variants vs Look-alikes, and selecting across groups
AI Curation Ranks a group and names every measure it used
Protected folders Folders whose files can never be selected
5. Review, then remove
Section titled “5. Review, then remove”Marking is not removing. Everything goes through Review Selection, which counts the batch, warns where a file would have no copy left behind, and then offers four ways out.
Review Selection The screen between marking files and removing them
The four operations Trash, delete, archive, copy — and which are reversible
6. History
Section titled “6. History”Every scan and every operation is written down, and most operations can be undone from a surviving copy — verified by content hash, months later, long after the OS Trash is gone.
History and restore Scan History, Operations History, and Review & Restore
Troubleshooting
Section titled “Troubleshooting”A folder was skipped — check that the OS granted DuoBolt access, under macOS Privacy & Security or Windows permissions.
Everything is slow — USB 2.0 enclosures and network mounts are simply slower. Antivirus scanning every read is the other usual culprit; allow-list DuoBolt during scans.
HEIC or AVIF show no preview on Windows — install the two free Microsoft Store extensions. The files are still found, hashed and grouped without them; see installation.
Fewer duplicates than expected — the default scan starts at 64 KiB and skips system and hidden files. What that excludes.
How matching works The engine, and what it does not claim
FAQ Hashing, safety, NAS, APFS clones, performance
Changelog Release notes and version history