Skip to content

macOS · iPadOS · iOS

Open the 2 GB JSON. Query it. Fix it.

JSON Scope opens the files other tools give up on. It maps the document, indexes it once, and stays interactive without ever loading it into memory. A 900 MB API dump, a 2 GB export, a multi-million-line log.

Everything happens on your device. No account, no network, nothing collected.

macOS 14.0 or later. iPadOS and iOS 17.0 or later. Universal Purchase.

api-export.json — 1.9 GB
4,182,904 items indexedmemory 34.1 MB
An illustration of the JSON Scope tree view showing a 1.9 GB file named api-export.json. Rows are indented by depth, each tagged with a type badge — curly braces for an object, square brackets for an array, “ab” for text, “12” for a number, “T slash F” for a boolean and a slashed zero for null. Containers show their child counts, including an array of 4,182,904 items. The status bar reports 34.1 megabytes of memory in use.

The problem

Every other tool fails at the size that matters

Backend dumps, database extracts and ML datasets routinely run to hundreds of megabytes. That is precisely where the usual options stop working.

  • Desktop viewers run out of memory

    Most load the whole document into a heap. One documents a 60 to 1 memory ratio, which puts a 500 MB file at 30 GB of RAM. In practice they cap out far below that.

  • Browser tools die well before a gigabyte

    A web viewer has to parse the file into a JavaScript heap, and it has to receive the file first. For proprietary data, uploading it is not an option at all.

  • The command line gives you no way to look

    jq streams happily, and then hands you a wall of text. There is no tree, no table, and no way to explore something whose shape you do not already know.

Free

A complete tool for everyday files

Free is not a demo. Files up to 50 MB, with the tree, the raw view, search, validation and Quick Look — no time limit and no nagging.

  • Files up to 50 MB

    Comfortably above every ordinary JSON file, so normal work is never interrupted by a prompt to buy something.

  • A virtualised tree

    Expands lazily at any depth. A container with ten million children opens without stalling, and every row shows its type and child count.

  • The raw bytes

    A byte view with correct line numbers for both Unix and Windows line endings. Select a node and the view reveals its exact span.

  • Find, with counts

    Search keys and values incrementally over the stream, with match counts and next and previous navigation. Cancellable at any point.

  • Validation that tells you where

    A syntax error reports the byte offset, the line, the column, the offending token and what was expected. Never a bare “invalid JSON”.

  • Quick Look in Finder

    Press space on a .json file and read a formatted, syntax-coloured preview without opening anything.

Pro

Scale, and the power tools

Pro removes the size limit and adds the querying, reshaping and editing that turn a very large file into an answer.

  • Any size at all

    Pro

    Tested against 2 GB fixtures. The file is streamed and indexed once, so resident memory stays flat no matter how large it gets.

  • Line-delimited logs

    Pro

    .jsonl and .ndjson, including event logs with tens of millions of lines.

  • Records as rows

    Pro

    An array of objects becomes a table with inferred columns. Types drive sorting and alignment, and ragged records leave empty cells rather than misaligned rows.

  • JSONPath and jq

    Pro

    Wildcards, recursive descent, slices, unions and filters. Or the jq syntax you already know, evaluated lazily so a query never materialises the file.

  • Schema and diff

    Pro

    Infer a JSON Schema 2020-12 for any node to document an undocumented API, or compare two documents structurally, keyed for objects and positional for arrays.

  • Edits that preserve your file

    Pro

    Change a value and save. Every byte you did not touch is written back verbatim, so whitespace, key order and numeric literals survive exactly. The save is atomic.

In use

On iPhone and iPad

The same engine and the same views, laid out for the screen you are holding. These are captures of the running app on a 4,000-record document.

  • JSON Scope on iPhone showing a nested document in the tree view, each row tagged with its value type.

    Open the 2 GB JSON.

  • The table view on iPhone, showing an array of objects as rows with inferred, sortable columns.

    Records as rows, types inferred.

  • The query bar on iPhone running a JSONPath expression, with the matching results listed below it.

    JSONPath and jq, on the whole file.

  • The raw view on iPhone showing the document’s source text with line numbers.

    The bytes as they really are.

  • The inspector on iPhone showing the selected node’s JSONPath, its raw value and its inferred schema.

    Path, schema and edits for any node.

On iPad

  • JSON Scope on iPad in a three-pane layout, with the document outline, the table view and the inspector side by side.

    Three panes on iPad.

  • The inspector on iPad showing a selected node’s path, raw value and inferred schema alongside the document.

    The inspector, with room to breathe.

Privacy

It cannot phone home, and that is enforced by the system

No network entitlement

On macOS the app ships without the network-client entitlement, so the operating system itself prevents it from opening a connection. It is not a promise; it is a capability the app does not have.

No third-party code

Zero runtime dependencies. No analytics package, no crash-reporting SDK, nothing that could collect anything on someone else’s behalf.

No account, ever

There is no sign-in and nothing to create. The App Privacy label reads “Data Not Collected”, because that is what the privacy manifest declares.

Your files stay put

Documents are read where they already live and are never copied elsewhere. Erase all data clears every cached index, recent, bookmark and query in one action.

The full detail is in the privacy policy.

Pricing

Free to use. Pro when you need the scale.

One purchase covers macOS, iPadOS and iOS. Both subscriptions include a 7-day free trial, and there is a one-time option if you would rather not subscribe at all.

Capabilities of JSON Scope, showing which are available in the free version and which require Pro.
CapabilityFreePro
Opening files
Open .json files up to 50 MBIncludedIncluded
Open .json files of any sizeTested against 2 GB fixtures. Memory stays flat regardless of file size.Not includedIncluded
Line-delimited .jsonl and .ndjsonIncluding files with tens of millions of lines.Not includedIncluded
Reading and navigating
Virtualised tree view with lazy expansionIncludedIncluded
Raw byte view with line numbersIncludedIncluded
Find keys and values, with match countsIncludedIncluded
Validate with precise line and columnReports the offending token and what was expected, not just “invalid JSON”.IncludedIncluded
Format and minifyIncludedIncluded
Copy JSONPath, value or subtreeIncludedIncluded
Querying and shaping
Table view with inferred columns, sort and filterArrays of objects become rows. Types drive sorting and alignment.Not includedIncluded
JSONPath queriesWildcards, recursive descent, slices, unions and filter expressions.Not includedIncluded
jq-style queriesIdentity, fields, indexing, slices, iteration, pipes, select, object construction, keys, length, map and to_entries.Not includedIncluded
JSON Schema 2020-12 inferenceNot includedIncluded
Structural diff of two documentsNot includedIncluded
Editing and exporting
Edit values and keys, then saveEvery byte you did not touch is written back verbatim. Saving is atomic.Not includedIncluded
Export any node as JSON or CSVNot includedIncluded
System integration
Quick Look previews in FindermacOS only.IncludedIncluded
“Query JSON File” action for Shortcuts and SpotlightNot includedIncluded