Tables & Fields

Tables are the core building blocks of your schema. Each table contains fields (columns) with types and constraints.

Creating tables

Via AI chat:

"Create a products table with id, name, description, price, and created_at fields"

Via UI: Click the + button in the schema panel to add a new table manually.

Field types

TalkingSchema supports standard SQL field types:

Category	Types
Text	VARCHAR, TEXT, CHAR
Numbers	INTEGER, BIGINT, DECIMAL, FLOAT, BOOLEAN
Date/Time	TIMESTAMP, DATE, TIME
Binary	BYTEA, BLOB
JSON	JSON, JSONB
Custom	Enum types you define

Field constraints

Each field can have one or more constraints:

Constraint	Description
Primary Key	Uniquely identifies each row
NOT NULL	Field must have a value
UNIQUE	No two rows can have the same value
Default	Value used when none is provided
Auto-increment	Automatically increments (for integer PKs)

Setting constraints via chat

"Make the email field on users UNIQUE and NOT NULL" "Add a default value of now() to the created_at field" "Make the id field auto-increment"

Editing tables manually

Click any table node in the ERD to open the table editor:

• Rename the table
• Add/remove/reorder fields
• Change field types and constraints
• Drag fields to reorder them

Naming conventions

TalkingSchema follows SQL best practices:

• snake_case for table and field names
• Plural table names (users, not user)
• id as the primary key name
• {table}_id for foreign keys (e.g., user_id, organization_id)
• created_at / updated_at for timestamps

Schema Changes Banner

Any time the AI modifies tables, a Schema Changes Banner appears at the top of the ERD panel. Click:

• "Keep all" — accept all changes
• "Undo all" — revert to the previous state