Creator and lead engineer

Brian MacDonald

I built Ownbox, and I run my own business on it.

Some called it a crazy idea. Time will tell.
Brian MacDonald, creator and lead engineer of Ownbox

Why I built Ownbox

It started with a belief: people and companies should own their data, or at least have a copy of it they can take with them and build on.

Then it became a need. My AI agents were working on too many computers, and some of them were working on my personal computer, with full access to all of my files. That was really concerning. They needed one place to work, and they needed to be contained.

And I didn't need another subscription, or another silo of data. That's the last thing any company really wants.

So I built a machine you own: one box where your AI agents and your team do real work together, on a server in your name, with the AI you already pay for. I think of it as freedom computing.

Before Ownbox

I grew up in Northern California in a family rooted in technology. I spent eight years at American Express, built and exited an e-commerce business in my twenties, played professional beach volleyball, and ran marketing operations and lead generation for dozens of companies. The thread never changed: build things that work, and build the marketing machine that grows them.

Those years taught me the hard truth about most founder-led companies: the founder isn't running a business, they're running a very demanding job. If you can't leave for a morning without the revenue noticing, you don't own the machine. You are the machine.

The build, counted

Before Ownbox, I designed, built and ran a production AI platform on my own, with real customers paying through Stripe. Every figure is counted from the repositories, not estimated.

  • 519,192lines of first-party code
  • 7,704commits
  • 3,737backend tests, all passing
  • 5months, one engineer

Read the engineering record

How I build

Moving fast is only safe with the right rails. These are the ones I hold every machine to.

  • Cost is metered, not discovered. A hard spend ceiling, checked before every AI request and recorded after it.
  • It asks before it acts. Anything that spends money or sends a message waits for a person's yes.
  • A hiccup never runs a job twice. A timeout is reconciled, never blindly retried.
  • Proven by breaking it. No test counts until the bug is put back and the test fails.