Dust
A Rust-based disk usage analyzer that shows which directories are eating your storage as an intuitive visual bar chart, instead of the cryptic nested output of 'du'.
🔗 Visit Dust
Description
The classic 'du' command can tell you how much space a folder uses, but reading its output to find the actual culprit — which subfolder, which file type — usually means piping it through sort and head and squinting at raw numbers. Dust does that analysis for you: point it at a directory and it shows the largest subdirectories and files as a simple bar chart, sorted by size, so the biggest space users are obvious in one glance.
Dust recursively traverses directories, highlighting the largest subdirectories first, with configurable output depth and entry count, multiple view modes (files only, directories only, or grouped by file type), support for various size formats (SI, binary, KiB/MiB), JSON output for scripting, and regex-based filtering. It's noticeably fast even on directories with many small files thanks to its Rust implementation, cross-platform, and free under the Apache-2.0 license, with 12.2k GitHub stars.
💬 Our review
The short version: instead of piping 'du' through sort and head to find what's eating your disk, dust shows you the answer directly as a ranked bar chart the moment you run it.
Against plain du, dust's value is entirely in presentation and speed — same underlying filesystem data, but instantly readable, and its Rust implementation handles directories with huge numbers of small files noticeably faster. Against ncdu, which offers an interactive drill-down TUI you navigate folder by folder, dust gives you the full picture in one shot without needing to explore — better when you just want the answer fast, while ncdu is better if you want to interactively hunt and delete. It's read-only, so pair it with your file manager or ncdu for actual cleanup, but as a fast first pass to find what's filling a disk, it's hard to beat and costs nothing under the permissive Apache-2.0 license.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Projet open source (licence Apache-2.0), aucune offre payante
Pros
Graphique en barres visuel et immédiat, contrairement à du
Très rapide même avec de nombreux petits fichiers (Rust)
Plusieurs modes d'affichage (fichiers, dossiers, par type)
Export JSON pour scripts et automatisation
12 200+ étoiles, licence Apache-2.0 permissive
Cons
Uniquement en ligne de commande, pas d'interface graphique
Nécessite une installation ou compilation séparée
Outil de lecture seule, ne permet pas de nettoyer directement