Free JSON Tree Viewer

JSON, the data format that won

Douglas Crockford published JSON on json.org in 2001 as a lightweight alternative to XML, drawing the syntax directly from JavaScript's object literals. The format was first standardised as an IETF Internet-Draft in RFC 4627 (July 2006), then revised as RFC 7159 (March 2014), and finally settled in RFC 8259 (December 2017), which is the current standard and is kept aligned with ECMA-404 (1st edition October 2013, 2nd edition December 2017). RFC 8259 added one significant requirement: the network-layer encoding must be UTF-8. Two complementary specs round out the ecosystem: RFC 6901 «JSON Pointer» (April 2013) defines the /store/book/0/title syntax for addressing nodes inside a document, and RFC 6902 «JSON Patch» defines a JSON document that describes changes to apply to another JSON document. JSON Schema (current IETF draft 2020-12) adds validation. Everything else built on top, JSONC for config-file comments, JSON5 for relaxed syntax, NDJSON for streaming, came after.

The six JSON value types (and what is missing)

RFC 8259 §3 defines exactly six value types. There is nothing else.

• object: unordered set of name-value pairs delimited by curly braces. Keys are always strings.
• array: ordered list of values delimited by square brackets. May contain mixed types.
• string: Unicode characters in double quotes. Single quotes are not legal JSON.
• number: IEEE 754 double-precision floating point. The spec does not guarantee arbitrary precision; numbers beyond Number.MAX_SAFE_INTEGER = 2⁵³ − 1 = 9 007 199 254 740 991 lose precision when parsed.
• true / false: boolean literals, lowercase, unquoted.
• null: the null literal, lowercase, unquoted. A key with the value null is different from a key that is absent.

What JSON does not have: a date type (convention is an ISO 8601 string with timezone), comments (use a separate metadata field), trailing commas, single-quoted strings, hexadecimal numbers, undefined, NaN, or Infinity. Any of those will make JSON.parse throw SyntaxError.

Strict JSON, JSONC, JSON5: which is which

Strict JSON (RFC 8259) is what this viewer accepts: no comments, no trailing commas, double-quoted keys, double-quoted strings. JSONC is a Microsoft convention shipped in VS Code that allows // line comments and /* ... */ block comments while keeping everything else strict; you see it in tsconfig.json, .vscode/settings.json, and the jsonc-parser npm package. JSON5 (2017, https://json5.org) is a superset that adds single-quoted strings, unquoted keys, trailing commas, hexadecimal numbers, IEEE 754 special values (NaN, Infinity), and both line and block comments; the json5 npm package has roughly 10 million weekly downloads. NDJSON / JSON Lines (https://jsonlines.org) is a streaming variant where each line is one independent JSON document, used by log shippers and big-data ingestion pipelines. Strip comments and trailing commas before pasting if your source is JSONC or JSON5.

Where a tree viewer actually earns its keep

• API response inspection. Stripe, GitHub, Slack, Twilio: every modern API returns nested JSON. Pasting into a tree viewer makes the shape obvious.
• Configuration files. package.json, tsconfig.json, composer.json, kubeconfig, and the long tail of cloud-provider JSON configs all become easier to audit as a tree.
• Database export inspection. MongoDB find() output, PostgreSQL row_to_json rows, ElasticSearch hit lists.
• Schema discovery. When an API has no documentation, the tree viewer is the documentation.
• Diff staging. Before running json-diff or jq between two files, inspect each in a viewer to spot obvious differences manually.
• Webhook payload inspection. Stripe, GitHub, SendGrid, and similar services all send JSON webhooks. Pasting the payload into a viewer is the fastest way to verify what your endpoint will receive.
• Sample data crafting. Build a test fixture by typing JSON and watching the tree update incrementally; catches missing braces before you save the file.

Common mistakes when working with JSON

• Trailing commas. {"a": 1,} is legal JavaScript but illegal JSON. JSON.parse throws SyntaxError: Unexpected token '}'. Strip them or switch to JSON5.
• Comments. // like this or /* like this */ are valid in JavaScript and JSONC but not in strict JSON.
• Big-integer precision loss. Twitter and Stripe IDs can exceed 2⁵³; JSON.parse("9007199254740993") returns 9007199254740992. Receive them as strings if precision matters.
• Confusing null with absent. {"foo": null} has foo defined as null; {} has foo undefined. The distinction matters for API contracts and database NULL handling.
• Duplicate keys. RFC 8259 says behaviour is undefined; in practice every JS engine keeps the last value. {"a": 1, "a": 2} parses as {a: 2}.
• Circular references. JSON.stringify throws TypeError: Converting circular structure to JSON. Flatten cycles, or use a replacer, or use a library like flatted.
• Date strings without timezone. "2026-05-12" is ambiguous; "2026-05-12T18:30:00Z" is unambiguous ISO 8601 in UTC. Always include the offset.

More frequently asked questions

How big a JSON file can this viewer handle?

Parsing 1 MB with JSON.parse takes around 10 ms in V8. Rendering 100,000 collapsed DOM nodes can freeze the browser for several seconds. In practice, this viewer is comfortable up to about 5 MB; beyond that, expect noticeable lag during render. For larger documents, use a streaming parser like stream-json in a Web Worker, or page the input into chunks.

Why does my big integer come back wrong?

JavaScript numbers are IEEE 754 double-precision floats. Integers up to 2⁵³ − 1 = 9 007 199 254 740 991 survive a round-trip exactly; beyond that, precision is lost. Twitter snowflake IDs (64-bit), Stripe customer IDs, blockchain transaction IDs, and many database surrogate keys exceed this limit. JSON.parse will silently truncate. The fix is either to receive IDs as strings (most modern APIs do this) or to parse with a library that supports BigInt.

What is JSON Pointer, and where would I use it?

RFC 6901 (April 2013) defines a path syntax for addressing nodes inside a JSON document: /store/book/0/title picks the title of the first book in the store array. JSON Pointer is the foundation of JSON Patch (RFC 6902), which expresses document deltas as arrays of pointer-and-operation pairs. The Kubernetes API server uses JSON Patch for incremental updates; so do GitHub's REST API and many other systems.

Does NDJSON / JSON Lines work in this viewer?

No. NDJSON is one independent JSON value per line, with no enclosing array. JSON.parse on the whole file throws because the second line begins where the first object ends. Wrap the entire content in [ and ] and replace each newline-between-objects with a comma, or paste one line at a time, or use a dedicated NDJSON viewer.

Is my JSON sent anywhere?

No. JSON.parse runs in your browser's JavaScript engine, the tree is rendered as DOM nodes locally, and no copy of your data is sent over the network. This is safe for API responses with secrets, customer data, or PII. Open the Network tab in DevTools and paste a sample: you will see zero outgoing requests during parsing and rendering.