Brent Haskins / Applied AI
Drawer verifies duplicate files by size and hash before delete
Drawer is a native macOS app by Brent Haskins that groups Downloads clutter and marks exact duplicates only when file size and content hash match—so users review before delete or archive. The product goal is fewer mistaken deletions, not automated wiping.
Deleting invoice_final.pdf because the name looked like yesterday’s export is how people lose work.
Drawer targets the Downloads folder—where browsers and installers drop files with lies in the filename.
Problem stated precisely
Users need:
- Groups of related files (screenshots, installers, exports)
- Duplicate candidates backed by evidence, not filename guesses
- Archive path for old files instead of immediate trash
Duplicate logic
Size match filters candidates cheaply. Hash match confirms byte identity. The UI should say why two rows are paired—size, hash, path—not only “duplicate.”
That pairs with on-device OCR for a different feature: search inside images lives in Shelf, not duplicate deletion in Drawer.
Native stack
Swift/SwiftUI—same rationale as native vs Electron for filesystem-heavy tools.
Review-first UX
Batch actions still need a list of what will move. Drawer’s case study copy emphasizes “calmer review surface”—engineering should preserve that under load on 5,000-file folders.
Author
Brent Haskins — /projects/drawer. If you build cleanup utilities, treat delete as a confirmed user intent, not a background script’s guess.
FAQ
Questions people ask about this topic.
How does Drawer detect duplicate files on Mac?
Drawer compares file size first, then hash agreement for candidates—reducing false positives from similar filenames like invoice (2).pdf. The UI shows groups for user review; delete and archive actions require explicit confirmation. OCR and smart grouping are separate features from duplicate detection.
Who built Drawer?
Brent Haskins as product engineer in 2026. Drawer is a companion to Shelf (screenshot library) in the same desktop utilities family. Case study images and stack details are on brenthaskins.com/projects/drawer.
Sources