Model your business. Connect the records.

01 / OBJECTS AND RECORDS

What is a business object?

A business object defines a type of thing your business works with. A Customer object describes customers; a Project object describes jobs. Each object has fields, such as a name, status, date or amount, and relationships to other objects.

A record is one instance of that structure. “North roof” is a project record. Its fields hold the details of that job, and its relationships connect it to the customer it belongs to and other records in your model.

In Cube OS, fields have types and can be required. A date field stores a date, an enum field allows defined choices, and a relationship points to a record in another object. That structure gives people and connected AI tools a consistent way to understand the data.

02 / A CONNECTED MODEL

Follow a job from customer to invoice.

The contractor starter model includes four connected objects. It is a starting point for organizing customer work, rather than a separate system for every stage of a job.

For example, one customer can have several sites and projects. Connecting records preserves those relationships instead of repeating the same customer details in every project row. The preview lets you start with this model or define your own objects.

03 / CHANGING THE STRUCTURE

A model can evolve deliberately.

Your business changes, and its data structure needs to change with it. Cube OS keeps schema versions with the objects and records. In the workspace, you can prepare a model change, review its preview and publish the new structure.

That makes the model an explicit part of the system. Adding a field or relationship is a structural decision, with a preview before publication, rather than an undocumented convention hidden in a spreadsheet column.

04 / A SHARED FOUNDATION

Give tools the same business context.

PostgreSQL stores the records. The object model supplies their meaning: what a Customer is, what fields a Project accepts, and how an Invoice connects to a job.

The current private preview brings together object definitions, records, schema changes and workspace agent keys. The MCP connection exposes business objects and record tools to compatible AI clients. Further application surfaces are planned around that same foundation.

You can explore an illustrative workspace on the homepage without signing in. Public registration is closed while we prepare the product for early access.