Pretty-print anything
echo '{"a":1,"b":[2,3]}' | jq .No arguments needed — bare jq . reformats compact
JSON into something a human can read.
jq
Slice, filter, and reshape JSON straight from the command line — no script file required.
$ curl -s https://api.github.com/repos/jqlang/jq | jq '.stargazers_count'
31842Why bother?
jq does for JSON what sed and grep do for
plain text: it lets a one-line command reach into a blob of data and
pull out, reshape, or filter exactly the part you need. It's a
single, dependency-free binary, so it drops into any pipeline —
shell scripts, CI logs, an API response you're poking at in a
terminal — without dragging in a language runtime.
Three things you can do right now
Pull out one field
jq '.user.login' event.jsonDot-paths walk into nested objects the same way you'd write them in JavaScript, minus the semicolons.
Filter a list
jq '.[] | select(.stars > 100)' repos.jsonselect() turns an array into just the elements
that pass a test — the workhorse of most real one-liners.
About this redesign
I like jq: it's small, fast, dependency-free, and I reach for it most days. But its homepage, jqlang.org, spends its first screen on a download panel and its next several screens on a changelog running back to 2013 — useful once you already know jq, but it never shows a newcomer what the tool actually does. There's no example anywhere on the page; you have to click through to the tutorial or manual before seeing jq run on real input.
This page moves a worked example above the fold, replaces the release history with three runnable recipes, and rewrites the pitch in plainer language. It links back to the real site for downloads, the manual, and everything else jq actually needs to ship — this is a better front door, not a replacement project.