GUIDE · VIEWS

See your data six ways —
then focus on the part that matters.

Every file opens as an interactive tree, but that's one of six views over the same document. And when a file is too big to take in at once, you can reframe any node as the root and work on just that subtree.

The six views

Switch from the View menu or with ⌘1–⌘6. They all read the same parsed document — no reload, no re-parse — and switching tabs keeps your place. A collapsible inspector column on the right summarizes whatever you click, in every view: rich previews for scalars, property lists for objects, and aggregations for arrays — a palette shows as merged swatches with counts, booleans as a true/false split, numbers as a histogram. In the tree, clicking a row selects it; the chevron — or double-clicking the row — expands and collapses.

Tree

The interactive collapsible tree — the default for JSON. Expand or collapse any node or whole subtree, toggle line numbers, and it stays smooth on huge files because it's virtualized.

Table

Array-of-records data as a sortable table, one column per field — a nested one-to-one object flattens into dotted columns (lead.name, lead.email), and right-clicking a header lets you flatten or collapse any column yourself. Cells keep their rich values — links, color swatches, sparklines — and clicking a [n]/{n} cell opens that container as the root. A profile strip under each header summarizes the column: a histogram and range for numbers, distinct count and top value for strings, missing rates everywhere.

Columns

Finder-style columns: select a node to open its children in the next column, and drive it all with the arrow keys. The detail panel shows the selection in full — including its hero, so a record's logo renders large and its series renders as a chart.

Graph

A node-link graph, paged lazily from the root outward, with a minimap and level-of-detail zoom — the fastest way to trace relationships across nested objects.

Raw

The document text in its native format — JSON pretty-printed, YAML as YAML, TOON as TOON — with themed syntax colors. The default view when you open a YAML or TOON file.

Schema

An inferred JSON Schema of the data, rendered in the same tree. It merges what it sees across every record, so a field present in most but missing in a few shows up as optional.

The table's profile strip answers “what does this column look like?” — click a cell for the full stats.

Schema and Graph need the whole document parsed in memory, so on a very large file that streams in, they may be unavailable until you narrow it down — which is exactly what rooting is for. The Table stays available, but sorting switches off when reading a cell from every row isn't affordable.

Open as root: focus on a subtree

Right-click any object or array and choose Open as root. QDev reframes that node as the document root and re-opens it as a standalone document — so the tree, search, jq, schema, and graph all now operate on just that subtree.

The right-click context menu on a tree node, showing “Open as root” and “Filter to this path”. — Right-click a container node → Open as root.

A breadcrumb trail appears while you're drilled in — root › teams › [0] — and every crumb is clickable to climb back out. The Exit control jumps straight back to the top.

The drill-in breadcrumb trail (root › teams › [0]) with a trailing Exit button. — The drill trail — click any ancestor, or Exit to return to the whole document.

● WHY IT MATTERS ON BIG FILES

Because rooting re-opens only that node's bytes, a subtree of a huge streamed file can come back small enough to regain Schema, Graph, and AI — capabilities the full file was too large for. It's the streaming-safe way to get full power on a slice.

Filter to this path

The same context menu offers Filter to this path, which runs a jq query for the node's path instead of re-rooting. It keeps you in the full document's context, but needs the file small enough to run jq over. Rooting is the better move when the file is too big for that; filtering is handy when you want a jq expression you can tweak.

Two more menu items move a node between views without re-rooting — View in Graph and View in Tree — and Copy pathgives you the node's location as jq, a JSON Pointer, or a dotted path.

Compare two documents

File → Compare… diffs the current document against a second one — paste it or open a file. Additions read green, removals red, and a changed value shows old → new on one line, with entry counts in the header. It opens on changes only; untick it to see every row in context. Objects compare by key and arrays by index, so an insertion mid-array reads as a run of changes rather than a shift. Both documents stay in the tab — nothing is uploaded. Comparison needs the current document parsed in memory, so it isn't available for very large streamed files.

Open a file and dig in.

Six views, one document, and a root you can move wherever you're looking.

Add to Chrome

Next: previews →

See your data six ways —then focus on the part that matters.

The six views

Open as root: focus on a subtree

Filter to this path

Compare two documents

See your data six ways —
then focus on the part that matters.