Seedling Overview

A schema-aware relational test data factory for Postgres and MySQL.

Seedling

Relational Test Data Factory

Seedling is a schema-aware test data generator. Point it at a Postgres or MySQL schema and it produces referentially-integer rows that respect every foreign key, unique constraint, and column type.

It auto-detects column types from your schema, maps them to built-in generators, and streams rows in dependency order (topological sort by FK). You can override any column with custom generators via a simple YAML DSL.

Key Features

• Schema introspection — Reads live Postgres and MySQL databases
• Auto-generated data — Column names and types map to realistic generators (email, name, phone, geography, finance, and more)
• Generator overrides — YAML-based DSL to customize per-column generation
• Deterministic output — Same schema + seed + count = byte-identical results
• Multiple output formats — SQL, CSV, JSON Lines, Parquet, direct DB insert, Postgres COPY
• Circular FK handling — Self-referencing tables and circular dependency loops
• Fast streaming — Parallel generation of independent table subgraphs
• Watch mode — Auto-regenerate on schema changes

Quick Example

# Introspect a live database
seedling introspect --db postgres://localhost:5432/mydb --output schema.yaml

# Generate 10,000 rows to a SQL file
seedling generate --count 10000 --output seed.sql

# Deterministic generation
seedling generate --count 50000 --seed 42 --output seed.sql

Database Support

Database	Introspection	Output
Postgres	Full (information_schema)	INSERT, COPY protocol, direct DB
MySQL / MariaDB	Full (information_schema)	INSERT, direct DB