What Is JSON? A Plain-Language Explanation

JSON shows up everywhere once you start looking for it. Configuration files. API responses. ChatGPT plugins. Browser developer tools. The export button in Google Analytics. If you have ever seen something that looks like {"name": "John", "age": 30} and wondered what it was — that is JSON.

You don't need to be a developer to understand JSON. It is essentially a structured way to write down information. Once you can read it, a lot of technical documentation becomes much less intimidating.

What JSON looks like

Here is a simple example — information about a person:

{
  "name": "Ji-hoon Kim",
  "age": 28,
  "city": "Seoul",
  "hobbies": ["cycling", "cooking", "reading"],
  "employed": true
}

Each piece of information has a key (name, age, city) paired with a value. The value can be text, a number, true/false, a list (array), or even another set of key-value pairs.

The 6 types of values in JSON

Type	Example	What it is
String	`"Seoul"`	Text, always in double quotes
Number	`28` or `3.14`	Integer or decimal, no quotes
Boolean	`true` / `false`	Yes/no value, lowercase only
Array	`["a", "b", "c"]`	A list of items in square brackets
Object	`{"key": "value"}`	Another set of key-value pairs
Null	`null`	Means "no value" or "empty"

Where you actually encounter JSON

Web APIs (the most common case)

When an app fetches your weather, shows today's exchange rate, or loads your social media feed — all of that data travels as JSON between the server and your screen. The developer console in any browser lets you see these JSON responses in action.

Configuration files

Programs store settings in JSON files. package.json for Node.js projects. settings.json in VS Code. manifest.json for Chrome extensions. If you have ever edited one of these, you've worked with JSON.

Data exports

Google Analytics, Shopify, and many SaaS tools offer JSON as a download format. It's machine-readable, which means you can import it into another tool without losing structure.

AI and chatbot APIs

If you use the ChatGPT API, Claude API, or any AI service directly, you communicate with it by sending and receiving JSON. The messages you send, the parameters you set, and the responses you get are all formatted as JSON.

The most common JSON errors — and how to fix them

Invalid JSON is the most common source of frustration when working with APIs or config files. Here are the mistakes that come up again and again:

Trailing comma. {"a": 1, "b": 2,} — that last comma before the closing brace is illegal in JSON. Easy to accidentally add when you delete a line.
Single quotes instead of double. {'name': 'John'} — JSON only accepts double quotes. Single quotes are not valid.
Comments in the JSON. // this is a note is not allowed in JSON. If you need comments, look at JSON5 or JSONC formats.
Mismatched brackets. Every { needs a matching }, and every [ needs a ]. Nested structures make this easy to lose track of.

If you are not sure what is wrong, paste your JSON into a formatter — it will tell you exactly which line the error is on.

JSON vs XML — which is which?

Before JSON became dominant, XML was the standard format for data exchange. You still see XML in some older systems, RSS feeds, and certain government APIs. Here is how they compare:

JSON

{
  "name": "Seoul",
  "pop": 9720000
}

XML

<city>
  <name>Seoul</name>
  <pop>9720000</pop>
</city>

JSON is shorter and easier to read. XML is more verbose but supports things JSON does not, like attributes and comments. For most modern web development, JSON has become the default.