Database

Flexible JSONB collections with access rules, realtime subscriptions, and no predefined schema.

How it works

Koolbase Database stores records as JSONB in PostgreSQL. You define collections — like tables — from the dashboard, and each collection can hold any shape of JSON data without a fixed schema.

Each collection has three configurable access rules — read, write, and delete — that control who can perform each operation. Rules are enforced server-side on every request.

Your fields are returned at the top level of each record. System metadata lives in a reserved $-prefixed namespace: $id, $createdAt, $updatedAt, $collection, and $createdBy (the creator's user ID, present only when the write was authenticated). Field names beginning with $ are reserved and rejected on write.

Key concepts

CollectionA named group of records — equivalent to a table. Created from the dashboard.

RecordA single JSON document within a collection. Has a UUID id and timestamps automatically.

Access RuleControls who can read, write, or delete. Options: public, authenticated, or owner.

RealtimeSubscribe to live record changes over WebSocket — created, updated, and deleted events.

GIN IndexPostgreSQL GIN index on the JSONB data column for fast field-level queries.

Records are shared across environments

Collections and records are project-scoped, not environment-scoped. Your dev and prod environments read and write the same records — environments isolate API keys, feature flags, remote config, and version policy, not data. For true dev/prod data separation, use separate projects.

In this section

Collections

Create and manage collections from the dashboard

Insert & Query

Add records and query with filters

Update & Delete

Modify and remove records

Access Rules

Control who can read, write, and delete

Realtime

Subscribe to live record changes over WebSocket

Database vs Storage

Use the Database for structured JSON data — user profiles, posts, settings, messages. Use Storage for files — images, documents, videos. Database records are not suitable for storing binary data or large blobs.