CommandsIntermediate4-6 min reference

Faker.js

Faker (@faker-js/faker) generates realistic fake data — names, emails, addresses, dates — so tests aren't littered with "test1", "test2". This is a quick lookup of the common generators and the one setting that matters most for stable tests (seeding). For when to use it versus fixtures, see Test Data Patterns.

Setup

import { faker } from "@faker-js/faker";
// locale-specific:
import { fakerDE as faker } from "@faker-js/faker";

Common generators

Call	Returns
`faker.person.fullName()`	"Jane Doe"
`faker.internet.email()`	"jane.doe@example.com"
`faker.internet.userName()`	"jane_doe"
`faker.phone.number()`	a phone number
`faker.location.streetAddress()`	a street address
`faker.date.past()` / `future()` / `recent()`	a Date
`faker.number.int({ min, max })`	an integer
`faker.string.uuid()`	a UUID
`faker.helpers.arrayElement([...])`	a random item
`faker.helpers.multiple(fn, { count })`	an array of generated items
`faker.commerce.price()` / `product()`	commerce data

Determinism — seed it

faker.seed(123);        // same sequence every run

Without a seed, a failing test isn't reproducible. Seed in a global setup (or per test) so CI and local match.

Tips

Use faker.helpers.unique(fn) cautiously (deprecated in newer versions — prefer your own uniqueness, e.g. append a counter/uuid).
Set a locale to match the system under test (validation, formatting).
Generate via factories that call Faker, not Faker calls scattered through tests.

Common mistakes

Not seeding → non-reproducible failures.
Relying on uniqueness without guaranteeing it (collisions cause flake).
Generating data that violates the app's validation (wrong locale/format).
Faker calls inline everywhere instead of behind a factory.

// Related resources