SkydiveAI coworkers that actually get work done
Promoted
DuoBolt
Duplicate file finder and look-alike photo culler macOS/Win
3 followers
Duplicate file finder and look-alike photo culler macOS/Win
3 followers
DuoBolt is a desktop application for macOS and Windows. It finds exact duplicate files by full BLAKE3 hashing, and groups look-alike photos with a separate engine that ranks them and explains the pick. All inference runs on the device; nothing is uploaded and no account is required.
DuoBolt prioritizes safety, performance, and clarity with a local-first architecture where all processing happens on your machine. The workflow is review-first, fully auditable, and reversible.
This is the 2nd launch from DuoBolt. View more
DuoBolt
Launching today
DuoBolt is a desktop application for macOS and Windows. It finds exact duplicate files by full BLAKE3 hashing, and groups look-alike photos with a separate engine that ranks them and explains the pick. All inference runs on the device; nothing is uploaded and no account is required.
DuoBolt prioritizes safety, performance, and clarity with a local-first architecture where all processing happens on your machine. The workflow is review-first, fully auditable, and reversible.
Free Options
Launch Team
Hey Product Hunt!
I'm Ruben, the solo developer behind DuoBolt 2.2 — a desktop app for macOS and Windows that finds duplicate files and, more importantly, helps you cull look-alike photos.
The problem: We all have folders full of bursts, exports, and re-saves. Deciding which copy to keep is tedious — so most of us just keep everything.
The solution: DuoBolt uses on-device AI to rank look-alike photos and explain why it picked one over another. It measures seven criteria: sharpness, resolution, exposure (and face exposure), format fidelity, eyes open, smile, and an aesthetic score. It tells you why it thinks one shot is better — but it never auto-deletes. The final choice is always yours.
Version 2.2 adds two powerful tools:
1. Compare two photos side-by-side with a curtain you drag across them, and a live histogram that shows you the light in each image — so you can see which version is sharper, better exposed, or keeps more detail.
2. Rules you write once. You can set an ordered list of rules — keep the RAW, keep the one with EXIF, keep the one with more pixels, keep the one inside a named folder, or set EXIF rules (ISO under 400, shutter faster than 1/125, aperture wider than f/2.8, at least 12 MP, has GPS). Apply the rules to one group or to every group at once.
What makes it different:
- All AI runs on your device. No uploads, no accounts, no hidden fees. Just your files, on your machine.
- RAW, HEIC and AVIF are read directly — not just the embedded JPEG preview.
- macOS Photos library support. DuoBolt scans the library directly, uses Photos' own names, shows you what Photos knows (favorites, edits, import dates), and removes through Photos' Recently Deleted.
- Built in Rust, solo, with no telemetry and no background network calls. Download is under 10 MB and the app is notarized.
- Free to try. 7 days of Pro, then the app reverts to a generous free tier. The full scan runs, the results are complete — only the amount you can remove in a single operation is capped.
- One-time purchase, 2 seats. $34.99 Standard, $64.99 Pro. No subscription.
Honest note: DuoBolt is not forensic recovery software. Restoring a file requires that a byte-identical copy still exists somewhere. And it doesn't do framing or composition scoring — it will tell you when two shots are genuinely different and refuse to choose between them, because that's a judgment about the photo's purpose, not a technical one.
I built this alone, in Rust, with a custom multi-platform Skia UI engine and BLAKE3 hashing. I'm happy to answer anything — the AI engine, the comparison curtain, the Photos library integration, or how to make a desktop app that stays out of your network.
DuoBolt 2.2: [duobolt.app](https://duobolt.app)
For the PH community, 20% off any license with the code PH20 (valid for one week).
What's the one thing you wish a duplicate finder could do, but never does?