MangoDisk

MangoDisk is an open source, safety-first disk cleaner, storage analyzer, and system toolkit built with Tauri 2 and a Rust core. It runs natively on macOS, Windows, and Linux, and ships the same cleanup engine both as a graphical desktop app and as a standalone mangodisk CLI.

The desktop app covers ten core workflows. Deep Cleanup scans system and app caches, browser data, package-manager downloads, IDE and Xcode data, container caches, project build artifacts (node_modules, Rust target, Gradle, Swift, Python, .NET, Godot, CMake outputs), and downloaded local AI models, grouping candidates by reclaimable space with smart recommendations. Large File Cleanup finds bulky installers, videos, and archives; Duplicate File Cleanup verifies exact duplicates by content rather than filename, always keeping at least one copy per group; and Disk Space Analysis renders a treemap or sunburst chart so you can drill down to the largest folders and files. On the maintenance side, MangoDisk adds a Privacy Cleanup (browsing history, cookies, recent items, clipboard traces), an Application Uninstaller that also removes leftover caches and settings, Startup Item Management, System Optimization, System Maintenance one-click fixes for common macOS/Windows problems, an Operation History log of every change, and a resource monitor with CPU, memory, and network stats available from the menu bar or system tray. Optional AI explanations (free daily quota, or bring your own AI service) describe what an item does before you act on it.

Safety comes first: scans are read-only by default, nothing is deleted without explicit confirmation, and every cleanup rule must be backed by vendor documentation, exclude personal files and protected paths, and pass validation on real systems. The full rule library and its revision history are public on GitHub. Unlike CleanMyMac X, DaisyDisk, or CCleaner, MangoDisk is fully GPL-3.0, covers macOS, Windows, and Linux from one codebase, keeps its rule set transparent, and offers a reproducible CLI for scripted use.