FileSweep
Cleaning a disk is scary because it is easy to delete the wrong thing. FileSweep first proves that files are identical, then proposes a plan that can be undone.
Two passes
First, files are grouped by size: files of different sizes cannot be duplicates. Then the candidates are hashed with SHA-256 by several workers at once. If a file changes while it is being read, that is detected and it stays out of the result.
A plan instead of deletion
Every move or trash action goes into a plan first. A safe move verifies size and hash, and across volumes it copies, verifies and only then removes the original. Everything is recorded in a history with undo.
A sorting machine
Files ride a conveyor, a size gate sorts them into lanes, a press with several heads stamps a SHA-256 fingerprint, and identical prints meet in one bin. Nothing reaches the shredder without passing the plan gate.
Screens
Under the hood
- A Go back end on Wails, an interface on React and TypeScript, SQLite with no native libraries.
- SHA-256 hashing with a worker pool and detection of changes during reading.
- Safe moves with size and hash checks; across volumes it copies, verifies, then deletes.
- History with undo, CSV export, Russian and English interface, light and dark themes.
- Builds and checks in GitHub Actions; the code is open.
Go · Wails · React · SQLite