Test data for enterprise systems

Ask the agent

Describe the data you want in plain English. The chat agent fetches, generates, masks, and seeds, streaming every step it takes.

Get started

Sign up, go from a prompt to your first dataset, then learn the core concepts.

Templates

The guardrails the agent obeys: 50+ field types, distributions, nested objects, and PII masking.

Connections

Where data is delivered: Postgres / MySQL / Mongo / MSSQL / Oracle / DB2, REST endpoints, or SAP OData V2/V4.

Scenarios

Reusable Python runs the agent writes for you. Trigger via REST or MCP, with live execution logs.

Agents & MCP

The same 25+ MCP tools the in-app agent uses, exposed to Claude, Cursor, or Copilot.

What DataMaker is

DataMaker is the agentic test-data platform from Automators. You describe the test data you need; the AI agent generates realistic, schema-valid records and delivers them straight into your SAP systems, databases, and APIs.

Generated by AI. Governed by you. Delivered into your systems.

The agent is the interface, but it can’t freelance. Every run is bounded by the templates you define (value ranges, schemas, which fields are writable, which are masked), targets the connections you’ve configured, and is fully audit-logged. That’s the point: AI you can let near your test systems.

From prompt to governed test data, in three steps

1. Prompt. Open a chat and ask: “fetch 200 customers from the SAP BusinessPartner service and seed them into the staging Postgres.” The agent plans and runs the steps, streaming each tool call as it goes.
2. Govern. Templates pin the shape and rules; masking keeps PII/PHI out of exports; scenario keys scope the agent to a project. You review what it did before anything touches a real system.
3. Deliver. Data lands in the target (a database table, a REST endpoint, or an SAP OData entity set), not a seed.sql you have to copy-paste.

What you can do with it

• Generate locale-correct, validated synthetic data (German USt-IDs that pass checksum, IBANs that pass MOD-97, IMEIs that pass Luhn).
• Fetch real records out of an existing system (typically SAP OData) for regression sets that match the shape of a real test case.
• Seed databases, APIs, or SAP entities directly.
• Mask PII at the template level so the agent never exports sensitive fields by accident.

Where to go next

New here? Start with a prompt: about two minutes from an empty workspace to usable data, then skim the core concepts.

Integrating into a pipeline? Jump to Scenarios (Python triggered from CI) or MCP (your own AI coding agent calling DataMaker tools).

Replacing SAP TDMS or a hand-rolled anonymiser? See Workflows → SAP regression and Workflows → Mask PII.