Steps and examples

Generate repeatable JSON test records

Start with a small object that has the shape your test expects. Choose explicit rules for important fields, then compare the generated records before exporting them.

How to use this tool

  1. Paste a template and inspect fields

    Paste a JSON object and select Inspect fields. An array of objects is accepted, but only its first object supplies the template; later objects are not combined.

  2. Set the record count and field rules

    Set Number of records and Repeatable seed. Use Add rule to choose Sequence, Example string, a numeric range, a text enum or another supported rule for an exact leaf-field path.

  3. Generate, review, then export

    Select Generate preview. Review warnings about retained template values. Copy full result or Download full result exports the retained result, even when its on-screen preview is shortened.

Input and expected output

Three users with explicit IDs and names

Workspace settings: Records: 3; seed: guide-demo; format: JSON. Add $.id as Sequence, start 101, increment 1; add $.name as Example string, prefix user_. Leave both null proportions at 0.

Input
{"id":0,"name":"source-name"}
Expected output
[
  {"id":101,"name":"user_1"},
  {"id":102,"name":"user_2"},
  {"id":103,"name":"user_3"}
]

These two rules are deterministic without random sampling. A fixed seed also makes random rules reproducible when the template and all settings stay the same. Generated names are sample strings, not realistic identities.

Limits and checks

Empty or incompatible templates
An empty array, scalar value or array containing non-objects is rejected. A rule must match an existing leaf field; it does not create a new field. The workspace allows up to 32 explicit rules.
Invalid ranges and oversized requests
Generate 1–1,000 records from at most 1 MiB of input. Reversed ranges, invalid dates and sequences outside the safe integer range are rejected. The 10 MiB output limit or complexity limits can stop generation before 1,000 records.

Understanding the result

Does this anonymize real data?

No. Keep template value rules preserve the original value, and precision-sensitive numbers default to preservation. Review those fields before sharing generated records; a seed is not an anonymization guarantee.

What changes when exporting CSV?

Top-level template fields become columns. Nested objects and arrays become JSON text inside cells. Null becomes an empty cell, so CSV cannot distinguish it from an empty string; formula-like cells need review before opening in a spreadsheet.