I built a tool to generate relationship-aware test data for Database (no seed scripts)

Dhanmoni Nath — Wed, 17 Dec 2025 10:29:22 +0000

I wanted to share a side project I’ve been working on and get some honest feedback from the community.

The idea came from a problem I kept running into - both at work and on my own side projects. Features would work perfectly in development, tests would pass, and everything would seem fine. But once the same code hit production, issues would start appearing. More often than not, the problem wasn’t the code - it was the data.

In development and testing, our datasets were usually small, clean, and unrealistic. Seed scripts helped initially, but as the application grew, schemas evolved, and relationships became more complex, those scripts became harder to maintain. Updating or rewriting scripts for every collection or schema change was time-consuming, easy to postpone, and often skipped entirely.

That frustration is what led me to start building Drawline.

What Drawline Does

Drawline is designed to make working with realistic test data easier:

You can connect an existing database or start with a fresh project
It automatically infers the database structure and relationships
You can visualize and tweak the schema using a visual interface
With one click, it populates a dev or test database with realistic, relationship-aware data
Schema versions can be tracked and restored, similar to how version control works for code

The goal is to help developers get meaningful data early-for testing, debugging, and demos - without having to write or constantly maintain seed scripts.

The product is still early, but I’m already using it for my own projects and iterating based on real usage.
If you’re curious, you can check it out at Drawline.

I’d Love Your Feedback

I’m sharing this not as a pitch, but to learn from other developers who’ve dealt with similar problems:

How do you currently handle test data for web applications?
Is this something you’d actually use, or do existing tools and scripts work well enough for you?
What would make a tool like this a hard “no”?

Thanks for reading, and I’d really appreciate any feedback or thoughts.

Forem: Dhanmoni Nath

Must read

Early data modeling for backend teams: migrations first or design the data layer in isolation?

Sauvik Nath ・ Feb 6

I built a tool to generate relationship-aware test data for Database (no seed scripts)