Open a messy JSON dump (or a whole NDJSON log) and read it as a sortable, filterable grid. Query it, flatten it, export to Excel. No jq, no online formatters, nothing leaves your Mac.
Universal · Apple Silicon + Intel · 8 MB · 7-day free trial
[ { "id": 1, "status": "OK", … }, { "id": 2, "status": "ERROR", … }, { "id": 3, "status": "OK", … }, { "id": 4, "status": "ERROR", … } ]
| id | status | note |
|---|---|---|
| 2 | ERROR | line1↵line2 |
| 4 | ERROR | =1+1 |
So you paste it into an online formatter you shouldn't trust with prod data, or wrestle jq until the array finally lines up. JSON Analyzer skips all of it: open the file, and arrays of objects are a spreadsheet: one you can sort, filter, query, flatten, and hand to Excel.
Files up to 50 MB open instantly as a dense nested grid, several at once as tabs. JSON Lines, NDJSON and log files open as one table, one row per line.
Press ⌘⇧F and type $.rows[?(@.status=="ERROR")]. Or just ask “rows where status is ERROR” in plain English. Stack Excel-style column filters on top.
Right-click → export the exact visible view to CSV, XLSX, or Markdown. Newlines, commas and quotes inside cells stay intact, which a clipboard copy can't do.
Structure shows through cell containment, not runaway indentation. Expand any node in place; 1k-row chunking keeps huge arrays fluid.
One click flattens any nesting into a single spreadsheet: every leaf becomes a column (user.address.city), arrays expand to rows. Query it like any table.
A documented JSONPath subset. $.rows[?(@.id>3)] filters a table to matching rows in place, and composes with sort, filters and editing.
Type what you want; the app writes the query. “Explain this JSON” summarizes an unfamiliar file's shape. Optional, uses your own Anthropic key, and you preview every request before it's sent.
Per-column value checklist with live counts, funnel indicators, and visible/total tallies. Column profiling on header hover: distinct, min, max, null share.
A real .xlsx with typed numeric cells. No dependency, no upload: what you see filtered on screen is exactly what lands in the file.
Edit scalars in place with type auto-detection and undo/redo. Save is byte-minimal: untouched bytes stay byte-identical to the original.
Check a document against a JSON Schema with inline errors on the exact cells. Going the other way, copy an inferred TypeScript type or JSON Schema from any file.
Compare an API response against yesterday's fixture side by side. Each tab keeps its own view, filters, query, and undo history.
Your JSON never touches a server. No sign-up, no cloud, no analytics — just a fast native app reading files on your own disk.
The two AI features are strictly opt-in: they use your own Anthropic API key, stored in the macOS Keychain, and show you the exact request (a capped outline of the file's structure) for approval before anything is sent. Skip them and the app never makes a network call at all.
Open it as a spreadsheet, filter to what broke, and export the proof, all in about ten seconds.