Configure Database

You are helping the user add a PostgreSQL database to their ChatGPT App using Supabase.

When to Add Database

• App needs to persist user data
• Data survives across conversations
• Multiple entities with relationships
• Query and filter capabilities needed

Workflow

1. Check Supabase Setup

Ask:

• "Do you have a Supabase project?"
• If no, guide them to create one at supabase.com

1. Gather Credentials

• Project URL
• Anon Key (public)
• Service Key (private, server-side)

1. Define Entities

For each entity, gather:

• Name (e.g., "Task", "Recipe")
• Fields and types
• Relationships to other entities

1. Generate Schema

Use chatgpt-schema-designer agent to create:

• SQL migrations in supabase/migrations/
• TypeScript types
• Query helpers

1. Generate Connection Pool

Create server/db/pool.ts.

1. Setup Local Development

Create Supabase config and start scripts.

1. Apply Migrations

   supabase start
   supabase db reset

Schema Requirements

Every table MUST have:

• id UUID PRIMARY KEY
• user_subject TEXT NOT NULL (for data isolation)
• created_at TIMESTAMPTZ
• updated_at TIMESTAMPTZ
• Index on (user_subject)

Environment Variables

DATABASE_URL=postgresql://postgres:[email protected]:54322/postgres

Query Pattern

Always filter by user_subject:

SELECT * FROM tasks WHERE user_subject = $1