I built a tool to generate relational SQL seed data with preserved foreign keys

Description and use

SeedSQL generates realistic test data that preserves foreign keys, speeding up local database setup and CI pipelines.

Problem

Hand‑writing mock SQL data for local testing is slow, and using LLMs breaks down when thousands of rows with strict relational constraints are needed.

Outcome

SeedSQL produces deterministic, foreign‑key‑aware seed data in seconds, eliminating manual effort and LLM limitations.

Workflow

• 1. Define database schema or upload it to SeedSQL.
• 2. Configure data types, ranges, and patterns for each column.
• 3. Run the generator to produce synthetic rows with correct foreign key mappings.
• 4. Export the output as INSERT statements for PostgreSQL, MySQL, SQLite, or as structured JSON.
• 5. Execute the statements against the local database or CI environment.

Evidence

Standard generators treat tables independently, creating orphaned relationships (Users → Orders).

Source

AI Limits: Prompting LLMs for large datasets gets slow, hits token caps, and yields non-deterministic outputs.

Source

Preserves Foreign Keys: Automatically maps parent-child primary keys across connected tables.

Source