Glossary

A short, opinionated glossary. The longer model is in Get started → Concepts.

Agent

The AI model that drives DataMaker via tool calls. Two flavours:

In-product agent — the chat in DataMaker’s UI. See Agent mode.
External agents — Claude / Cursor / Copilot, talking to DataMaker via the MCP server. See MCP server.

Connection

A configured target system: database, REST endpoint, or SAP OData service. Knows the URL, auth method, and (for OData) the entity sets it discovered. See Connections.

Field

A single column in a template. Has a name, a type, and type-specific options. ~50 built-in types; see Field types.

Generation

One call to dm.generate() (or the equivalent REST/UI). Produces N rows from a template. Counts toward your hourly row quota.

MCP

Model Context Protocol — the JSON-RPC stdio protocol AI agents use to call tools. DataMaker ships an MCP server with 25+ tools. See MCP server.

Project

A logical grouping inside a workspace. Templates, connections, scenarios, chats, and sets all live inside a project. Most teams have one project per environment.

Run

A single execution of a scenario. Has a status (running / done / failed / cancelled), a duration, and a log. See Scenarios → Logs & retries.

Scenario

A Python script that orchestrates a sequence of DataMaker actions. Lives in a project, runs on hosted Python workers, triggerable from UI / API / MCP / chat. See Scenarios.

Sensitive

A field property. When sensitive: true, DataMaker substitutes a fake at generation, refuses agent-driven exports, redacts in logs, and writes an audit entry on every export. See Templates → Sensitive fields.

Set

A named, versioned snapshot of rows. Created with dm.save_set() / POST /sets; reloaded with dm.load_set() / GET /sets/{name}. Used for regression workflows.

Template

A reusable schema definition: a list of fields, each with a type and options. Templates are versioned. See Templates.

Workspace

Your team. Has a billing plan, members, API keys, and one or more projects.