SaaS

One PostgreSQL backend for the product you are shipping.

A small SaaS product can keep its accounts and relational data in one CubiBase project. Users sign up through CubiBase Auth; customers, plans, and the records they own live in the same database. The API role returns a row only when the user column matches the JWT subject, if you set that column up.

Model the account on a column you control

For a project tracker, a tasks table can carry owner_id as text. Set owner_id as the user column at create time. Ada signs up through /auth/v1/signup, and her access token’s subject is her user id. When it calls GET /rest/v1/tasks, the API role is limited to rows whose owner_id equals that subject. Your web app still renders the product. CubiBase answers the auth and data calls.

Shared records need a column that matches every subject who should see them, or a policy you are willing to design for. CubiBase’s generated policy is equality on one text column. It does not generate organization trees, role matrices, or billing entitlements.

A single product database

Use one project for one product. If a second product must not see those tables, give it another project and another database. CubiBase does not split a project across regions, and it does not attach an uptime SLA to this setup.

Questions

Can each of my customers get a separate CubiBase database?

You can create a project per customer. That is a stronger boundary than a user column, and it is more projects to manage. For a typical multi-tenant SaaS app, one project database plus a user column is the model CubiBase generates today.

Does CubiBase bill my customers?

No. CubiBase has not published its own prices, and it does not run your subscription billing.

Start with one project.

Add another project when the next application needs its own database.

first request
GET /rest/v1/customersapikey: cb_pub_…Authorization: Bearer <JWT>200 · { "data": [ … ] }