Generate JSON Mock Data for Unit Test Fixtures

This is the literal output for count 2, seed 42, Sequential IDs off. Anyone who sets the same seed and count gets exactly this — which is what makes it safe to check in.

Controls: count 2, seed 42, Sequential IDs off

[
  {
    "id": "ab732cc0-86ce-45f2-ad8f-78e93ffbe1ce",
    "index": 1,
    "name": "Grace Davis",
    "email": "grace.davis@mock.dev",
    "phone": "+1-555-7711",
    "age": 36,
    "status": "active",
    "address": { "street": "550 Cedar Blvd", "city": "Ogdenville", "zip": "10345" },
    "score": 47.1,
    "createdAt": "2024-01-17",
    "active": false
  },
  {
    "id": "c89687c4-06f1-4c4a-bcdc-cf1c51c8732c",
    "index": 2,
    "name": "Carol Smith",
    "email": "carol.smith@example.com",
    "age": 21,
    "status": "pending",
    "score": 30.4,
    "createdAt": "2023-11-15",
    "active": false
  }
]

Turning Sequential IDs on swaps the UUID for id: 1, 2, 3…. Note the non-id fields differ from the previous example even at the same seed — that is expected, because the id mode shifts the internal random stream.

Controls: count 3, seed 42, Sequential IDs ON

[
  { "id": 1, "index": 1, "name": "Grace Davis", "age": 28, "status": "pending", "active": false },
  { "id": 2, "index": 2, "name": "Frank Wilson", "age": 46, "status": "suspended", "active": false },
  { "id": 3, "index": 3, "name": "Bob Taylor", "age": 37, "status": "inactive", "active": true }
]
// (fields trimmed for brevity — every record still carries the full shape)

The fixed shape is a feature here: because the field names never change, a single factory keeps tests terse. Generate the base array once, import it, and spread overrides per test.

// fixtures/users.json was generated at seed 42, count 10
import users from './fixtures/users.json';

const makeUser = (overrides = {}) => ({ ...users[0], ...overrides });

test('suspended users cannot log in', () => {
  const u = makeUser({ status: 'suspended', active: false });
  expect(canLogin(u)).toBe(false);
});

test('active adult passes', () => {
  const u = makeUser({ status: 'active', age: 30, active: true });
  expect(canLogin(u)).toBe(true);
});

Random UUID ids change with the seed, so do not assert on them. The index field is always 1..n in order, which makes a dependable positional assertion.

import users from './fixtures/users.json'; // seed 42, count 5

test('records are in stable order', () => {
  expect(users.map(u => u.index)).toEqual([1, 2, 3, 4, 5]);
});

test('id is opaque — never assert its literal value', () => {
  // GOOD: shape check
  expect(typeof users[0].id).toBe('string');
  // BAD (brittle): expect(users[0].id).toBe('ab732cc0-...')
});

If your unit under test only needs id, name, and status, generate the full records and drop the rest with json-key-filter so the fixture is minimal and the diff is readable.

Step 1 — generate count 5, seed 42 here.
Step 2 — paste into json-key-filter, keep only: id, name, status

Result:
[
  { "id": "ab732cc0-...", "name": "Grace Davis", "status": "active" },
  { "id": "c89687c4-...", "name": "Carol Smith", "status": "pending" }
]

Older versions of this page implied you could paste an example object and the tool would infer fields. It cannot. The generator ignores any input and always emits the fixed person shape. To reach a custom shape, generate the base array and reshape it with json-key-filter and json-key-renamer.

Sequential IDs does more than swap the id field — turning it on means the UUID generation step no longer consumes random numbers, which shifts every subsequent field's value. So the same seed produces different name, age, status, etc. between the two id modes. Pick the id mode first, then settle on a seed.

Names come from a pool of 10 first × 10 last = 100 combinations, and emails are derived from the name plus 5 domains. At counts above ~100 records, repeated names and even duplicate emails are statistically certain. If your test asserts email uniqueness, post-process or use a smaller count — the generator does not guarantee unique emails.

The free plan caps generation at 500 records; the input is clamped and a request above it shows an upgrade message. Pro raises the cap to 10,000 records. There is a hard internal safety ceiling of 100,000 records regardless of plan.

Output is always pretty-printed with 2-space indentation; the UI has no formatting control. For a minified single-line fixture, run the output through json-minifier. To re-pretty-print with a different indent, use json-prettifier.

The on-screen preview truncates past 5,000 characters and appends a …(truncated) marker so the page stays responsive. The Copy and Download actions always contain the complete output — verify by checking the record count shown above the preview.

Phones are always +1-555-NNNN — the 555 exchange is the reserved fictional-number range, so a fixture can never collide with a real US number. If a test parses phone numbers into a specific national format, account for this single shape.

Dates are generated in the year range 2020–2024 with day 01–28 (so February never overflows). If a test needs a createdAt in the future or near now, override that field after generation rather than expecting the generator to produce it.

All domains are non-routable test domains. A test that accidentally tries to send real mail will fail to deliver rather than spam a real address — a safety property, not a bug. Keep your test transport mocked regardless.