JSON Tools for AI Developers · Runs locally

For AI developers JSON Workbench

Natural language generation · Schema · Zod · OpenAPI · Function Calling · MCP · 30+ tools

Batch mock

List pages, paginated APIs, and light load tests need a set of same-shaped objects, not a single record.

Paste a Schema or sample on the left, set a count, click Schema → Mock, and the right pane shows an array.

Types are inferred from the Schema. enum and format are followed when possible; complex oneOf needs a human tighten.

Batch data is generated in the browser. It never hits your backend.

Start small, then grow

Generate 3 records, check the shape, then go to 20 or 50. Hitting 100 first and finding a missing field is more work.

A Schema is more stable than a bare sample

With JSON Schema, required and type keep the output in line. With only a sample, inference is as optimistic as single Mock.

How to use it

  1. Paste a JSON Schema or sample and set a count (1 to 100).
  2. Click Schema → Mock.
  3. Spot-check a few records: required present, types not drifting.
  4. Copy the array into the frontend, Storybook, or a test fixture.
  5. Validate one record against the Schema so this batch is still usable.

When to use it

  • Ten fake records while the list page has no API.
  • Same-shaped objects for a pagination component.
  • Build a local payload batch before a light load test.
  • Accept a Schema: look at a few edge values in the batch on purpose.

Keep in mind

  • The UI allows 1 to 100 records. Loop or script for more.
  • This is not a distributed load tester. It only builds data.
  • The Schema is not uploaded.

FAQ

What's the max per run?

The UI allows 1 to 100. Loop yourself or use a script for more.

Can I use it without a Schema?

Yes — paste a JSON sample as the shape. A Schema keeps constraints tighter.

When should I use single Mock instead?

Use single Mock to confirm the shape, then batch.

Is the Schema uploaded?

No.

Is every record different?

Values change; the shape stays the same. Don't assume emails are globally unique unless you dedupe later.

Can I set a random seed?

There is no seed input. Save one output as a fixture if you need it reproducible.

How is oneOf generated?

Complex combinations may follow a common branch. Spot-check, and tighten the Schema if needed.

How does this work with the Table page?

Paste the batch into Table and scan whether columns line up.

Related tools

JSON Mock JSON Schema JSON Schema validate Back to the Workbench →