VeilDB
Database anonymization and secure sharing across developers
Many years ago, I was working as a developer in an e-commerce company.
One day, we received an urgent Slack message: customers were being charged again.
Someone had imported a production database dump into staging without disabling cron jobs. The project was subscription-based.
Billing ran again.
~$150,000 charged.
Weeks of refunds. Damage control. Reputation hit.
That incident stayed with me.
The Pattern I Kept Seeing
As I grew from developer → tech lead → CTO, I kept seeing the same problem:
Teams share production databases in unsafe ways.
Common approaches:
“We have a script to trim data.”
“Developers clean it manually.”
“We have rules - people just need to follow them.”
In reality:
Production data handling depends on the discipline.
And discipline breaks under pressure.
Even after we built internal tooling, someone from an offshore team once imported production data into staging and broke a third-party search integration.
We fixed it quickly.
But I realized something:
This problem is systemic.
I Tried to Find a Tool
Back in 2020, I looked for a proper solution.
What I found:
Enterprise-only platforms
Stack-specific tools
Nothing universal
Nothing simple
So I built one internally.
Later, I decided to turn it into an open-source project.
Introducing VeilDB
VeilDB is an open-source database anonymization and safe-sharing tool.
The idea is simple:
Instead of trusting humans to clean production dumps, automate the process.
Core principles:
The original database is never modified.
Dumps are processed in isolation (Docker).
Sensitive fields are masked based on configurable rules.
Developers only download sanitized dumps.
Access is permission-based and token-secured.
Architecture:
Service (UI for rules & permissions)
Agent (runs on your server, processes dumps)
Client (developers download processed backups)
That’s it.
No magic.
Just removing human error from the workflow.
Why Open Source?
Two reasons:
Trust - If we’re talking about data processing, the logic should be transparent.
Adoption - This problem exists in thousands of teams.
Where It Is Now
Core functionality works.
Documentation is ready.
Initial version released.
Early feedback coming in.
Still early.
Still polishing.
Why I’m Posting Here
I’m curious:
How do you handle production database sharing?
Do you trust scripts?
Do you automate masking?
Or is it still “be careful”?
Also:
If you’ve built B2B infra tools before, I’d love advice on distribution.
Open source first?
Or push towards managed SaaS faster?
If you’re interested, here’s the repo:
https://github.com/veildb-tech/service
Happy to answer any questions.
Comment