Random JSON Generator
Define fields, nest objects/arrays, weight custom lists, set ranges, add derived formulas, and link multiple tables — instantly, in your browser. 100% private.
What is a Random JSON Generator?
A random JSON generator creates realistic mock data for development and testing. Instead of manually writing test fixtures, you define the structure — field names, types, ranges, even nested objects and arrays — and the tool generates thousands of realistic records instantly: names, emails, UUIDs, dates, coordinates, and much more.
Mock data is essential for frontend development (building UI before the API is ready), unit and integration testing (reproducible test fixtures), load testing (filling a database with realistic data), API prototyping (demoing endpoints without production data), and presentations (realistic-looking demos without exposing real user data).
Available Field Types
60+ types across nine categories, plus five structural types — enum (a weighted custom list), object (nested fields), array (repeated primitives or objects), formula (derived from sibling fields), and reference (joins to an earlier table).
| Category | Example types | Sample output |
|---|---|---|
| Identity | fullName, username, uuid, mongoId, password | Arjun Patel · a1b2c3d4-e5f6-4… |
| Contact | email, phone, phoneIntl, url, avatarUrl | arjun.patel42@example.com |
| Address | streetAddress, city, state, zipCode, latitude | 4271 Main St, Surat, 394210 |
| Number | int, float, price, percent, rating | 87.43 · 4.5 ★ · 12% |
| Date & Time | date, datetime, pastDate, futureDate, unixTimestamp | 2024-11-03 · 1735689600 |
| Text | word, sentence, paragraph, slug, hexColor, emoji | cache-matrix-42 · #3f9be2 |
| Web / Tech | ipv4, mac, userAgent, jwtToken, httpStatusCode | 192.168.4.17 · 404 |
| Commerce | sku, isbn, creditCardNumber, companyName, jobTitle | SKU-7F2A91 · Mehta Labs |
| Structural | enum, object, array, formula, reference | { tags: ["a","b"] } |
Ranges, Weighted Lists, Nesting and Null Injection
Number fields (int, float, price, percent, rating, unixTimestamp) accept a min/max and a choice between a uniform or normal (bell-curve) distribution. A Custom List field takes one value per line and an optional weight (active:7) so, for example, 70% of generated records can land on "active". Object and Array fields let you nest a sub-schema one level deep — useful for an address object or an items: [...] array on an order. Every field also has an optional null-injection percentage, so you can test how your code handles missing data without hand-editing the output.
Formula Fields and Linked Tables
A Formula field derives its value from sibling fields already defined above it, using {fieldName} placeholders — {firstName} {lastName} joins two strings, while {price} * {qty} is evaluated as arithmetic (detected automatically from whether the template contains + - * /). The expression is parsed by a small built-in evaluator, not eval(), so nothing but arithmetic ever runs.
The tabs above the field list let you add more than one table to the same workspace — useful for relational-shaped test data. Give a field the Reference type and point it at a field on an earlier table (e.g. an orders table's user_id pointing at a users table's id) and it samples an actual generated value from that table, so the two arrays are genuinely joinable — not just two unrelated sets of random data. Tables generate in order, left to right, so a table can only reference ones before it.
Exporting and Reusing a Schema
The output isn't locked to JSON — the Download row converts the active table's records to CSV, SQL INSERT statements, NDJSON, or YAML, reusing JsonKit's existing converters (nested objects/arrays are flattened with dot notation for CSV/SQL). The Open in… row hands the current output straight to the JSON Formatter, Schema Generator, TypeScript, CSV, or YAML tools without a copy-paste round trip. If you already have a real JSON or CSV sample, Import… infers field names and best-guess types from it instead of building a schema by hand. And Share encodes the whole workspace — every table, its fields, record count and seed — into a URL, so a teammate opening the link gets the identical multi-table generator, with nothing saved on a server.
Seeding Random Data in Go Tests
// go get github.com/brianvoe/gofakeit/v6
import "github.com/brianvoe/gofakeit/v6"
type User struct {
ID string
Name string
Email string
Age int
}
func randomUser() User {
return User{
ID: gofakeit.UUID(),
Name: gofakeit.Name(),
Email: gofakeit.Email(),
Age: gofakeit.Number(18, 80),
}
}
// Generate a slice of random users
func randomUsers(n int) []User {
users := make([]User, n)
for i := range users {
users[i] = randomUser()
}
return users
}Mock Data in JavaScript Tests
// npm install @faker-js/faker
import { faker } from '@faker-js/faker';
function randomUser() {
return {
id: faker.string.uuid(),
name: faker.person.fullName(),
email: faker.internet.email(),
age: faker.number.int({ min: 18, max: 80 }),
active: faker.datatype.boolean(),
city: faker.location.city(),
createdAt: faker.date.past().toISOString(),
};
}
// Generate 10 random users
const users = Array.from({ length: 10 }, randomUser);
console.log(JSON.stringify(users, null, 2));