JSON validator, formatter and repair

How to validate and format JSON

1. Paste your JSON into the editor, open a file, or drag one in.
2. Press Validate (or Ctrl/Cmd + Enter). If something is wrong, you get the line and column, what was expected, and how to fix it. Go to error jumps right there.
3. Beautify to indent it with 2 spaces, 4 spaces or tabs, or Minify to strip every optional space. Then copy it.

Repair broken JSON

Press Repair to turn "almost JSON" into valid JSON in one click, and see exactly what changed. It removes trailing commas and comments, swaps single and curly quotes for double quotes, quotes unquoted keys, turns True, None, NaN and undefined into JSON values, fixes numbers like .5, +1 and 0x1F, escapes line breaks inside strings, adds missing commas, colons and closing brackets, and strips Markdown code fences and JSONP wrappers. JSON that was saved as a string, like "{\"a\":1}", is unwrapped. Undo with Ctrl/Cmd + Z.

Validate JSON against a schema

Open JSON Schema and paste a schema (draft 4, 7, 2019-09 or 2020-12). Validate then checks the shape of your data too: required properties, types, formats like email and date-time, enums, ranges and patterns. Each problem shows its JSON path and jumps to the exact line. It works in Bulk validate as well, so you can check a whole folder of files against one schema.

Compare two JSON documents

Open Compare and paste two versions, like an API response before and after a change, two config files, or expected and actual test output. The diff is by meaning, not by text: key order, spacing and formatting never count, 1 equals 1.0, and "\u00e9" equals "é". You get a list of added, removed and changed values with their JSON paths, each linked to its line. Turn on Ignore array order when lists are sets.

Validate many JSON files at once

Open Bulk validate and drop in a batch of files or a whole folder, like fixtures, translations, configs or an API dump. Each file gets a pass or fail with the line, column and fix for the first error, plus warnings like duplicate keys. Click a file to open it in the editor at the error, and copy a plain-text report to share. Thousands of files work, checked in the background in your browser.

Errors it explains

navaja json follows the JSON standard (RFC 8259) strictly, and spots the mistakes people actually make:

• Trailing commas after the last item, like [1, 2,].
• Single quotes and unquoted keys, like {'a': 1} or {a: 1}, which are JavaScript, not JSON.
• Missing commas between properties or items.
• Comments, which JSON doesn't allow.
• Python and JavaScript values like True, None, NaN or undefined.
• Invalid numbers: leading zeros, hex, .5, 1., a leading plus.
• Broken strings: a missing closing quote, a real line break or tab inside a string, a bad escape like \x.
• Invisible characters pasted from web pages, like non-breaking spaces, and curly quotes from documents.
• Mismatched brackets: a ] closing an object, or a bracket that is never closed.

It also warns about valid JSON that will likely bite you later: duplicate keys, where most parsers silently keep only the last value, and numbers too large for JavaScript, which JSON.parse would round. Formatting here never changes your numbers.

Questions