DBSEED

Context-aware AI-powered database seeder for generating realistic test data.

Features

• 🎯 Interactive Mode - Guided prompts for easy data generation
• 🔍 Schema Introspection - Automatically reads your database structure
• 🔗 Foreign Key Support - Generates data respecting relationships
• 🎲 Smart Data Generation - Context-aware fake data using Faker.js
• 💾 Flexible Output - SQL file or direct database insertion
• ⚡ Fast - Built with Bun and TypeScript
• 📦 Standalone Binary - No runtime dependencies

Installation

curl -fsSL https://raw.githubusercontent.com/db-toolkit/dbseed/main/install.sh | bash

Or download the binary from releases.

Quick Start

Interactive Mode

dbseed

Follow the prompts to:

1. Enter database connection string
2. Select tables to seed
3. Choose number of rows
4. Select output method (file or database)

Generate Command

# Generate 1000 rows and insert into database
dbseed generate -c "postgresql://user:pass@localhost:5432/db" -n 1000

# Generate SQL file
dbseed generate -c "postgresql://user:pass@localhost:5432/db" -n 500 -o seed.sql

# Seed specific tables
dbseed generate -c "postgresql://user:pass@localhost:5432/db" -t users posts -n 100

Introspect Schema

# View all tables
dbseed introspect -c "postgresql://user:pass@localhost:5432/db"

# View specific tables
dbseed introspect -c "postgresql://user:pass@localhost:5432/db" -t users posts

Options

Generate Command

• -c, --connection <string> - Database connection string (required)
• -t, --tables <tables...> - Specific tables to seed
• -n, --rows <number> - Number of rows per table (default: 100)
• -o, --output <file> - Output SQL file (omit for direct DB insert)
• --ai - Use AI for context-aware generation (coming soon)

Introspect Command

• -c, --connection <string> - Database connection string (required)
• -t, --tables <tables...> - Specific tables to introspect

Supported Data Types

• Strings, Text, Email, URL, Phone
• Integers, BigInt, Decimal, Float
• Boolean, UUID, JSON/JSONB
• Date, DateTime, Timestamp, Time
• Arrays, Binary/Blob
• Foreign Keys

Column Name Recognition

DBSEED intelligently generates data based on column names:

• email → realistic email addresses
• phone, mobile → phone numbers
• first_name, last_name → person names
• company → company names
• address, city, country → location data
• price, amount → currency values
• created_at, updated_at → timestamps
• And many more...

Development

# Install dependencies
bun install

# Run in dev mode
bun run dev

# Build binary
bun run build

Requirements

• Bun runtime (for development only)

License

MIT

Contributing

Contributions welcome! Please open an issue or PR.