KOSUNI KOREA
Korean News in English
Technology

Building Tables with AI: A Step-by-Step Guide for Developers

Published October 16, 2026 · 3 views
Building Tables with AI: A Step-by-Step Guide for Developers

What You'll Learn in This Lesson

Last week, we designed the database structure needed for a dating app.

In this lesson, we'll implement that design into an actual database.

Previously, developers had to write SQL queries manually,

but now, with Cursor AI and Supabase MCP, most of the work can be done collaboratively with AI.

This lesson will cover:

  • Requesting SQL from AI

  • Reviewing generated SQL

  • Creating tables

  • Verifying the results

  • Modifying and regenerating as needed

Let's get started.


Starting Implementation After Design

In our previous lesson, we designed these tables:

  • users

  • profiles

  • likes

  • matches

  • chat_rooms

  • messages

  • notifications

We'll now create each one in Supabase.


First, Requesting from AI

We ask Cursor AI to generate SQL for our tables with a clear request:

Generate SQL based on Supabase PostgreSQL for a dating app database.

Create the following tables:

- users
- profiles
- likes
- matches
- chat_rooms
- messages

Also include Primary Keys and Foreign Keys.

Providing context about your project and required tables helps AI produce more accurate SQL.


Reviewing AI-Generated SQL

AI usually generates correct SQL,

but it's important to review before executing directly.

Check these key points:

  • PRIMARY KEY configuration

  • FOREIGN KEY relationships

  • UNIQUE constraints

  • NULL allowance settings

  • DEFAULT values

SQL review is a standard practice in professional development.


Creating the users Table

We'll start with the users table to store member information.

This table will include:

  • User ID

  • Email

  • Creation date

Only essential data is stored here, as Supabase Auth handles login functionality to avoid duplicate information.


Creating the profiles Table

Next comes the profiles table for user details:

  • Nickname

  • Gender

  • Age

  • Location

  • Bio

  • Profile image URL

This table connects 1:1 with the users table.


Creating the likes Table

The likes table implements the 'like' feature:

  • from_user_id

  • to_user_id

  • created_at

When requesting from AI, specify that a user shouldn't be able to like the same person multiple times by adding UNIQUE constraints.


Creating the matches Table

Matches are created when two users 'like' each other:

  • user1_id

  • user2_id

  • matched_at

Request AI to prevent duplicate matches for the same user pair.


Creating the chat_rooms Table

Chat rooms are created after a match:

  • room_id

  • match_id

  • created_at

Linking chat rooms to match information simplifies management.


Creating the messages Table

The messages table stores most data:

  • room_id

  • sender_id

  • message

  • created_at

AI recommends appropriate column lengths and data types.


Creating the notifications Table

We'll set up a basic notification structure:

  • New likes

  • New matches

  • New messages

This lesson focuses on creating the foundational structure.


Executing SQL

After reviewing AI-generated SQL,

run it in Supabase's SQL Editor.

Once complete, open Table Editor to verify all tables were created successfully.


Asking AI About Errors

If you encounter errors during execution:

  • Typo in column names

  • Foreign key issues

  • Data type mismatches

Ask AI to analyze the error with a request like:

Analyze this SQL error and provide solutions.

AI will explain causes and fixes in detail.


Verifying Created Tables

In Table Editor, check:

  • Number of tables

  • Column names

  • Data types

  • Default values

  • Relationships

This ensures the implementation matches our design.


Improving Structure with AI

After creating tables, you can ask AI to suggest improvements:

What performance optimizations exist in this structure?

Or:

Which columns should have indexes added?

AI will analyze your project and propose enhancements.


Modifying with AI

As development progresses, you might need new columns:

For example, to add MBTI data:

Create SQL to add an mbti column to the profiles table.

AI automatically generates ALTER TABLE statements for these changes.


Good Practices

When modifying databases, follow this sequence:

  • First, backup

  • Review SQL

  • Test in a staging environment

  • Apply to production

Always review AI-generated SQL before execution.


Next Steps in Development

With this database complete,

Your app's foundation is ready for implementation:

Registration

Login

Profile setup

Now you can begin building actual features.


Next Lesson Preview

The next lesson will cover creating a Next.js project and starting dating app development.

We'll initialize the project, apply Tailwind CSS, and begin front-end development.


Conclusion

This lesson demonstrated using Cursor AI with Supabase MCP to create database tables for a dating app.

With both design and implementation complete,

We're ready to start building user interfaces in the next lessons, moving toward a fully functional dating application.

FAQ

Can I trust AI-generated SQL without review?

No, always review AI-generated SQL before execution to ensure correctness and security.

What is Supabase MCP?

Supabase MCP (Managed Cloud Platform) is a cloud-based database service that simplifies database management for developers.

Korean Source

This article is an English localization of a Korean-language report from BlogScore. Original: [바이브코딩 강좌] 20강. AI와 함께 테이블 생성하기