What Is dust? The Beginner's Guide to the Modern du Alternative

CodexFuse Team · 2026-11-04T21:55:00.984Z · 11 min read min read

What Is dust? The Beginner's Guide to the Modern du Alternative

To be clear up front: this is not about the grey powder on your shelves, and not about the DUST memecoin on BNB Smart Chain either — what is dust, in the Linux world, is a free, open-source command-line tool (the dust CLI) by Andy Boot (bootandy) that answers one question better than anything before it: "Where did all my disk space go?" Its official tagline is "du + rust = dust. Like du but more intuitive," and the name really is just du + rust. Run one command, and dust draws your disk usage as a tree with bar graphs — the biggest directories float to the top, the largest one is colored, and you can see the space hog in seconds. If your box is a VPS hosting plan whose disk fills up silently between deploys, dust is the five-minute tool that stops the guessing.

What Is dust, Exactly?

dust is a non-interactive du alternative written in Rust. Think of this page as a dust explained walkthrough for total beginners: you give dust a directory, and it prints an instant overview of which subdirectories and files are eating your disk — as an ASCII tree with percentage bars, sorted biggest-first, with the largest entries colored. The creator's "why" for the project is one sentence: "Because I want an easy way to see where my disk is being used."

Three things make dust different from classic du. First, there is no sort or head pipeline: dust "will list a slightly-less-than-the-terminal-height number of the biggest subdirectories or files and will smartly recurse down the tree to find the larger ones," so you never need -d or -h just to get a readable answer. Second, the output is visual: bars show each entry's share of its parent, and the bar colors encode the tree hierarchy — shades of grey tell you which parent a subfolder belongs to. Third, it is opinionated about what to show: as the README puts it, "Dust simply Does The Right Thing when handling lots of small files & directories. Dust keeps the output simple by only showing large entries."

The project is Apache-2.0 licensed, has gathered over 12,500 GitHub stars since its 2018 debut, and shipped version 1.2.7 in November 2026. It is a single binary called dust — note that the crates.io and most distro packages are named du-dust (the binary stays dust).

Why du Alone Isn't Enough to Find Space Hogs

The classic way to hunt a space hog with du is a pipeline: du -d 1 -h | sort -h, or du -sk * | sort -nr. It works, but it is three mental steps — estimate, sort, then squint at a flat list that tells you nothing about nesting. If you are checking disk space on Ubuntu at 2 a.m. because a deploy failed, that pipeline is friction you do not need.

dust collapses those steps into one visual answer. The FreshPorts port description makes the case well: "Unless your intent was just to pipe du's output, dust is more likely to give you the utilization information you were actually looking for." Where du answers "how big is everything," dust answers the question you actually asked: "what is big, where, and how big relative to its parent?" The largest subdirectory is colored, its bar shows its percentage share, and dust recurses smartly so the real culprit — say, a 30 GB node_modules three levels deep — shows up without you drilling down manually.

How Do You Install dust? (dust install)

Install routes depend on your system, and the package name is the gotcha: on most distros and on crates.io the package is called du-dust, while Arch, Homebrew, and Scoop use plain dust. The binary is always dust. All of these were verified against the official README and repology this week.

curl -sSfL https://raw.githubusercontent.com/bootandy/dust/refs/heads/master/install.sh | sh
cargo install du-dust
brew install dust            # macOS and Linux
sudo dnf install du-dust     # Fedora
sudo pacman -S dust          # Arch Linux
scoop install dust           # Windows

The README also lists conda install -c conda-forge dust, mise use -g dust, and deb-get install du-dust, and version 1.2.7 ships ready-made Debian packages (du-dust_1.2.7-1_amd64.deb) in its GitHub release assets if you would rather not use cargo. One packaging caveat: the Snap build can only see files inside /home, so prefer another route for system directories. Confirm it works with dust --version, then point it at your home directory and watch it work.

How Do You Run dust for the First Time?

Open a terminal and type dust in any directory — that is the whole dust tutorial. But to make the output click, here is the five-minute first run, adapted from the project's own demo.

  1. Run dust with no arguments in your home folder. dust walks the tree and prints the biggest entries with bars and percentages, automatically limited to just under your terminal height.
  2. Read the biggest colored entry first. The colored entry is the largest subdirectory; its bar shows its share of the parent. In the README's own demo, a 72K webapp shows 32K in src — "the bar jumps to 44%. Half of the webapp is in src."
  3. Limit the depth with -d. dust -d 2 shows two levels of subdirectories — handy when the tree feels deep.
  4. Limit the count with -n. dust -n 5 shows only the five biggest items, perfect for a quick triage.
  5. Reverse the view with -r if you prefer smallest-first, and pass a path (dust /var/log) to scan anywhere.

The official usage cheat sheet is worth keeping: dust <dir>, dust -p (full paths), dust -s (apparent size), dust -n 30 (30 entries), dust -d 3 (3 levels), dust -r (reverse), dust -e "\.png$" (regex filter). One flag to remember by heart: -F shows only files — "finds your largest files" — which is the fastest way to spot one giant log or database dump.

Which dust Flags Should You Learn First?

You do not need all twenty-plus flags. Learn these six and you will handle 90% of disk-space hunts: -d (depth), -n (how many entries), -F (files only), -D (directories only), -r (reverse order), and -p (full paths so output is copy-pasteable). For narrowing a scan, add -z 10M (only files bigger than 10 MB) and -e/-v (regex include/exclude).

A few flags exist for special cases: -s shows apparent file size instead of disk blocks, -x stays on one filesystem, -i hides hidden files, -c disables colors for monochrome terminals, -b drops the bars entirely, -R switches to a screen-reader-friendly layout, and -j prints JSON for scripting. --collapse=node-modules keeps a huge folder collapsed instead of recursing into it. Persistent preferences live in a config file at ~/.config/dust/config.toml (or ~/.dust.toml), using the long flag names — for example reverse=true. Note one naming trap: -f counts files by inode instead of measuring disk space — it is not a filter, and -i is "hide hidden files," not "ignore by name" (that is -X).

dust vs du vs ncdu: Which Should You Use?

These tools do not compete so much as cover different moments. Here is the honest breakdown, built only from what the projects document about themselves.

ToolWhat it isBest moment to use it
duThe GNU coreutils classic that "estimates file space usage." Finding hogs needs a pipeline like du -d 1 -h | sort -h.Already installed everywhere; scripting; one-off checks inside other commands.
dust"du + rust = dust. Like du but more intuitive." Non-interactive tree + percentage bars, biggest-first, no sort or head needed. Visualizes only — it cannot delete anything.Instant answer to "what is eating my disk," especially on a server you SSH into.
ncduThe ncurses interactive du browser. Interactive browsing in a text UI — and it can delete files.When you want to browse into folders and delete the junk interactively.
dua-cliRust disk-usage analyzer with both CLI and interactive modes; "faster in some cases."Massive filesystems where parallel scanning pays off.

The practical rule, echoed across the guides: run dust for the instant overview, switch to ncdu when you want to browse and delete interactively. dust's own README draws the line against ncdu and baobab explicitly — "you have no idea where the largest files are. Finding this out via these other tools is not trivial whereas Dust will show the large file clearly in the tree hierarchy." And as net7.be puts it, dust "is not an interactive tool and behaves more like a replacement to the traditional du command."

What dust Can't Do (Read This Before You Blame It)

dust visualizes disk usage — it does not clean, delete, or free a single byte. Its flag list contains no delete option at all; ncdu and gdu are the ones with delete keys. If dust shows you the 40 GB log file, removing it is still your job.

Two more limits, both shared with du itself. dust walks the file tree like du does, so it does not see disk space held by deleted-but-still-open files (that is a lsof problem, not a dust problem). And dust tells you nothing about disk health — usage versus failing hardware are different jobs, and this article's sibling guide covers the health side. Treat dust as the fastest possible answer to "what is using it," then reach for the right tool for everything else.

Where Should You Point dust First?

Three places earn their keep on almost every machine. First, your home directory: run dust ~ and the -F flag will surface the forgotten 8 GB ISO or the video export hiding in Downloads. Second, project folders: dust -d 2 in a web project finds the bloated node_modules or build cache in one glance — --collapse keeps it tidy if you want the summary only.

Third, the system paths that fill servers: dust -n 10 /var/log for runaway logs and dust /var/lib/docker if Docker's disk usage has been creeping. Database servers deserve their own check too — checking a PostgreSQL database's size directly tells you whether the database or the logs are the real problem. For a first sweep of the whole machine, sudo dust -n 20 / with -x (stay on one filesystem) gives the complete picture in seconds.

Sources

Every fact in this guide was checked against a live source on 2026-11-04. Quotes marked "verbatim" are copied exactly from the project's own documentation.

Frequently Asked Questions

Is dust a replacement for du?

Yes, for interactive use it is meant to be: dust gives the same disk-usage information as du but as a sorted tree with bar graphs, with no sort or head pipeline. du itself still wins in scripts and on machines where you cannot install anything, since it ships with coreutils everywhere.

Can dust delete files to free up space?

No. dust visualizes usage only — it has no delete flag at all. If you want to browse folders and remove the junk interactively, use ncdu (ncurses, deletes) or gdu (press d to delete), which are built for that.

How do I install dust on Ubuntu or Debian?

The safest route is cargo install du-dust, or download the du-dust_1.2.7-1_amd64.deb package from the GitHub releases page — the binary is always named dust. The exact apt package name differs between Debian and Ubuntu packaging sources, so avoid guessing it and use cargo or the .deb.

Why is the package called du-dust but the command is dust?

The project's name is dust, short for "du + rust." The cargo crate and most distro packages use du-dust to avoid colliding with other packages, but every install method ends with a binary named dust — so you type dust at the terminal.

dust vs ncdu — which should I use?

Use dust when you want the fastest possible answer to "what is eating my disk" — one command, one glance. Use ncdu when you want to browse into the directory tree interactively and delete files as you go. They pair well: dust for the instant overview, ncdu for the cleanup session.

Featured on Twelve Tools Hunt for Tools Featured on LaunchBuff Launch Llama Tools Launch your MCP Top Web Development Companies in the US - OnToplist.com Designer Listings Listed on SeeWhatNewAI