JSON Formatter

Chrome extension that helps you view and explore JSON API responses.

Features

• Fast, even on very long JSON pages
• Dark mode
• Syntax highlighting
• Collapsible trees, with indent guides
• Clickable URLs
• Negligible performance impact on non-JSON pages (less than 1 millisecond)
• Works on any valid JSON page – URL doesn't matter
• Buttons for toggling between raw and parsed JSON
• Parsed JSON is exported as a global variable, json, so you can inspect it in the console (now working again!)

Some JSON documents for testing it on: https://callumlocke.github.io/json-formatter/

Installation

Option 1 (recommended) – Install it from the Chrome Web Store.

Option 2 – Install it from source (see below).

Development

Clone repo and run bun install.

Commands:

• bun run build - single build
• bun run watch - same but watch-driven

You can install dist as a local, unpacked extension in Chrome with developer mode enabled.

FAQ

How does it detect JSON?

This turns out to be a complex thing to get right in a bulletproof way. In most cases it's based on the Content-Type header but in some cases it's necessary to inspect the 'page' strucure and see if it looks like a JSON endpoint. This is designed to work as fast as possible with no perceivable impact on browsing.

Why are large numbers not displayed accurately?

This is a limitation of JavaScript and therefore a limitation of JSON as interpreted by your web browser.

• Anything over Number.MAX_SAFE_INTEGER (2^53 - 1 or 9007199254740991) is adjusted down to that number.
• Anything below Number.MIN_SAFE_INTEGER (-2^53 + 1 or -9007199254740991) is adjusted up to that number.
• Extremely precise floating point numbers are rounded to 16 digits.

It's not JSON Formatter doing this, it's the native JSON.parse in V8. JSON Formatter shows you the parsed values, exactly the same as what you'll see if you fetch your JSON from any web application.

If your API endpoint really needs to represent numbers outside JavaScript's safe range, it should quote them as strings.

Why are object keys sometimes in the wrong order?

What you see in JSON Formatter is a representation of the parsed object/array. It's the same order you'll get with Object.keys( JSON.parse(json) ) in JavaScript.

Historically, the JavaScript standard explicitly stated that object keys can be iterated in any order, and V8 took advantage of this by moving numeric string keys (like "1" or "99999") to the top to facilitate a small performance optimisation. This V8 implementation detail has since become standardised.

But I just want to see exactly what the server spits out

For now, your best option is to just use the "Raw" button to see the raw JSON. This is what the server sent. The "Parsed" buttons represents what you'll get from JSON.parse.

In future JSON Formatter might switch from using JSON.parse to a custom parser (if performance allows) in order to detect when a value has been 'changed' by parsing and show an appropriate warning.

Roadmap

Before v1.0:

• Rewriting & modernising for improved JSON-detection logic and configurability (underway)
• Better rendering architecture
• Show more request info including headers where applicable

For v1.0 or later (tentative):

• Customizable themes
• Support broken JSON, with error highlighting
• Better copy & paste
• Copyable JS property chains
• A way to easily see when parsed values are different from the raw JSON (numbers outside range, key reordering)
• A way to use jq syntax (or similar) in the browser
• A way to use JSON Formatter on pasted-in JSON