The catalog

GET /v1/catalog answers what a workspace's data looks like. Read it before asking anything of the data: it names every object and field the API, conditions and Liquid use, with what each means.

• Objects: the built-in ones (store, profile, customer record, order, line item, product, product variant, coupon, coupon code, event) and the workspace's published object types. Each field has its type (string, number, boolean, datetime, array, object or money), a description, whether it can be empty, its values when they're a fixed set, and whether conditions can filter and group by it. Nested fields are named by their path, e.g. shippingAddress.countryCode.
• Relationships: how objects connect, e.g. an order has lineItems, a profile has customers (its customer records, one per store).
• Events: every event name the workspace has, with its property keys and the types their values had. When a name was first and last seen, its events and how many happened over time are under /v1/event-types.
• Custom fields: the workspace's own fields on profiles, read and written as customFields.<key>.
• Computed traits: facts computed over related objects, such as a profile's orders count and total spent (per currency, never converted).

Built-in objects are described from the same schemas the API answers with, so the catalog always matches the API. It holds what exists, never how much: it's cheap to read every time.