Disk Usage by File Type
Paste a file listing with sizes — `du -ah`, `find -printf '%s %p\n'`, `ls -lR` or a size,path CSV — and see which file types eat the space, as a sorted bar chart with size, share and file count per extension or category. Nothing leaves your browser.
About this tool
Desktop disk-space analyzers answer one question first: which file types are eating the
drive? This tool gives you that same ranked breakdown from a listing you already have — a
du dump pasted from a server, a find run, an ls -lR capture, or a two-column CSV export.
Every line's size is added to its file extension (or to a broad category such as video or
documents), and the result comes back sorted biggest-first with the size, the share of the
total, a bar and the file count for each type.
It runs entirely in your browser on the text you paste. No folder is opened, nothing is uploaded, and no agent has to be installed on the machine that holds the files — which is why it works for a listing you captured over SSH from a box you can't install software on.
Worked example
Paste this du -ah output:
4.0K ./src/app.js
2.0M ./assets/hero.png
1.0M ./assets/logo.png
18M ./media/clip.mp4
8.0K ./README.md
With the defaults (group by extension, sort by size, binary units, 32-character bars) you get:
Disk usage by extension — 5 file(s), 21.0 MiB total
.mp4 18.0 MiB 85.7% ████████████████████████████████ 1 file(s)
.png 3.0 MiB 14.3% █████▍ 2 file(s)
.md 8.0 KiB 0.0% ▏ 1 file(s)
.js 4.0 KiB 0.0% ▏ 1 file(s)
Switch Group by to Category and the same listing collapses to video 18.0 MiB 85.7%,
images 3.0 MiB 14.3%, documents 8.0 KiB, code 4.0 KiB — the fastest way to see whether a
disk is full of media, build artifacts or source.
Listings it reads
| Command | Example line |
|---|---|
du -ah / du -a | 4.0K ./src/app.js |
find . -type f -printf '%s\t%p\n' | 4096 ./src/app.js |
ls -l, ls -lRh | -rw-r--r-- 1 me staff 1024 Jan 3 10:11 notes.md |
| CSV / TSV export | 4096,src/app.js |
Size suffixes are 1024-based, the way du -h and ls -lh print them, so 4.0K is 4096 bytes
and 1.5MiB is 1572864. A bare number is bytes. Thousands separators (1,234,567) are
accepted. Lines with no readable size — total 48 headers, tree art, prose — are skipped and
counted in a note under the chart, so you can tell whether the tool understood your paste.
Limits and edge cases
- Up to 20000 sized files per run. Past that the tool stops with an error rather than
truncating silently. For a whole disk, run
du -aone folder deeper and paste per folder, or feed itfind's output for the subtree you care about. - Folder rows are skipped by default.
du -aprints a line per folder whose size already includes everything beneath it; counting those would double-count the same bytes. A line counts as a folder when the path ends in a slash, whenls -lmarked itd, or when another pasted path sits inside it. Turn the checkbox off if your listing is files-only and you want every line counted verbatim. .tar.gz,.tar.bz2,.tar.xz,.tar.zststay whole instead of being counted as.gz/.xz. Any other double extension is counted by its last part.- Names with no usable extension —
README,Makefile,.gitignore,report v1.2 final— are grouped under(no extension)rather than guessed at. - Sizes are apparent file sizes, whatever your listing reported. Block-level allocation, sparse files, compression and hard links are not modelled: two hard links to the same file appear twice if the listing lists them twice.
- Percentages are of the pasted listing, not of the disk. Paste a partial listing and the shares are partial too.
- Beyond the top rows (15 by default) the remaining types are folded into one
(other N)row that still counts toward the total, so the percentages always add up.
FAQ
What exactly do I paste in?
Any listing where each line carries a size and a path. The usual sources are
du -ah, find . -type f -printf '%s\t%p\n', ls -lR (or ls -lRh), and CSV/TSV exports
with a size column and a path column in either order. Suffixed sizes (4.0K, 18M,
1.5MiB) are read as 1024-based, matching du -h; bare numbers are bytes.
Why is my total smaller than what `du -sh` reports?
Two common reasons. First, folder rows are skipped by default so their contents are not
counted twice — the total is the sum of the files in your listing. Second, du reports
space allocated on disk (rounded up to block size, and it counts a hard-linked file once),
while a find -printf '%s' listing reports apparent file size. Both are legitimate
numbers; this tool simply adds up whatever your listing said.
Can it group by something broader than a file extension?
Yes — set Group by to Category and extensions are rolled into images, video, audio,
documents, archives, code, data, executables, fonts and other (plus (no extension)). It's the
quicker read when you want to know whether a drive is full of media or of build output.
Unrecognised extensions land in other, and you can always switch back to the per-extension
view to see exactly which ones.
How do I export the numbers?
Set the output shape to CSV for name,bytes,size,percent,files rows that paste straight
into a spreadsheet, or JSON for a structured object with total_bytes, total_files,
skipped_folders, ignored_lines and a groups array. Both keep the exact byte counts
alongside the human-readable sizes. The Download button under the result saves whatever is
currently shown.
Can I get a picture rather than text?
Choose the Colored SVG bar chart output shape. It returns SVG source with one colored bar
per file type (categories get a fixed colour each), plus size and share labels and a hover
tooltip per bar. Copy it into a README, a ticket, a slide or a report — it's a plain text
<svg> element, so it stays sharp at any size and needs no image hosting.
Are my file names or sizes uploaded anywhere?
No. The analysis runs as WebAssembly inside your browser tab on the text in the box; the page makes no request with your data and nothing is stored. The same computation is available offline through the command-line tool if you'd rather keep the listing on the machine that produced it.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool disk-usage-by-filetype "12582912 ./media/intro.mp4
2097152 ./img/hero.png
1048576 ./img/logo.png
40960 ./src/app.js
8192 ./README.md"New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/disk-usage-by-filetype/?listing=12582912%09.%2Fmedia%2Fintro.mp4%0A2097152%09.%2Fimg%2Fhero.png%0A1048576%09.%2Fimg%2Flogo.png%0A40960%09.%2Fsrc%2Fapp.js%0A8192%09.%2FREADME.md&group_by=extension&sort_by=size&order=desc&top_n=15&units=binary&chart_width=32&skip_folders=true&ignore_case=true&format=chartMachine-readable descriptor: tool.json — title + parameters JSON Schema for agents.