What Is JSON? A Plain-English Guide

JSON in one sentence

JSON (JavaScript Object Notation) is a text format for representing structured data as nested key/value pairs and lists. It is language independent — Python, Java, Go, Rust, PHP and every browser can read and write it — but its syntax was borrowed from JavaScript, which is where the name comes from.

Because it is plain text, JSON can travel over HTTP, sit in a log line, live in a database column or be saved as a .json file. That portability is the entire reason it replaced heavier formats for most web APIs.

The syntax rules

JSON is deliberately tiny. There are only two container types and four scalar types, and the grammar fits on a single page.

• Objects are wrapped in curly braces and hold unordered key/value pairs: { "id": 1 }
• Arrays are wrapped in square brackets and hold ordered values: [1, 2, 3]
• Keys must be strings in double quotes — single quotes are invalid
• Values may be a string, number, boolean, null, object or array
• Members are separated by commas, and a trailing comma is not allowed

{
  "id": 4821,
  "name": "Ada Lovelace",
  "active": true,
  "score": 98.5,
  "manager": null,
  "roles": ["admin", "editor"],
  "address": {
    "city": "London",
    "postcode": "NW1 6XE"
  }
}

The data types you can use

JSON is not a JavaScript object

This is the single most common misunderstanding. A JavaScript object is a runtime value in memory; JSON is a string. JavaScript object literals allow unquoted keys, single quotes, comments, functions, undefined and trailing commas. JSON allows none of those.

{
  id: 1,          // unquoted key + comment
  name: 'Ada',    // single quotes
  tags: ['a',],   // trailing comma
}

{
  "id": 1,
  "name": "Ada",
  "tags": ["a"]
}

Where you will meet JSON

• REST and GraphQL API request and response bodies
• Configuration files such as package.json, tsconfig.json and composer.json
• Structured logging, where each log line is one JSON object
• Document databases and JSONB columns in PostgreSQL
• Browser storage, message queues, webhooks and CI pipeline outputs

Reading JSON comfortably

Real-world JSON arrives minified — one long line with no whitespace — because that saves bandwidth. Machines do not care, but people do. Formatting adds indentation and line breaks so nesting becomes visible, and a tree view lets you collapse the parts you are not interested in.

Common mistakes

Best practices

• Pick one key casing convention (camelCase or snake_case) and keep it across the whole API
• Prefer null over an empty string when a value is genuinely absent
• Keep arrays homogeneous — one shape per array makes consumers far simpler
• Send minified JSON over the wire, and format it only when a human is reading it
• Validate untrusted JSON before parsing it into typed structures