Built to run. Built to move.
People ask what HollywoodOS runs on. Fair question. Here's the honest answer.
HollywoodOS runs in production on AWS today but is not permanently locked in.
Real services, handling real data right now. Not a reference architecture. Not a diagram waiting for funding. And the part that makes it work doesn't live in anybody's cloud — it lives in the system itself.
Where it runs
HollywoodOS runs on managed cloud infrastructure. A managed SQL database holds the record. A container runtime runs the system. The kind of proven, standard services a serious institution already trusts and already knows how to reason about.
I'm not going to lay out the wiring on a public page. The shape of the deployment is something I'll walk you through in a demo, under whatever agreement fits. What matters before that conversation is simpler: it's real, it's running, and it sits on infrastructure you already understand.
No exotic hardware. No science project. The novel part of HollywoodOS is the governance — not the metal it runs on.
Hardware-agnostic by design
Most systems get welded to whoever's cloud they were born in. The vendor's database. The vendor's services. The vendor's lock-in. Moving later means rebuilding — so nobody ever does.
HollywoodOS was built the other way around. The rules — how it decides, what it allows, how it keeps its record — live in the system itself. Not in any one vendor's metal. The cloud underneath is where it runs. It is not what makes it work.
AWS is the hardware of record. It is not a dependency.
And that's not a claim I'm asking you to take on faith. The part of the system that talks to AI already runs across more than one provider. The portability is already proven, in the place it's hardest to pull off. If a better way to run it shows up tomorrow, the system is built to move to it.
That's the difference between built to last and built to lock you in. HollywoodOS is built to last.
What it takes to run it
Here's the practical answer for whoever has to sign off on it.
It's cloud-hosted. There's nothing to rack in your building. No on-prem requirement. No server room, no hardware to buy and babysit, no appliance to install. The managed services underneath are ones your people most likely already work with.
What an institution brings is its own field and its own authority — the rules for its level, and the people who hold the power to approve things. HollywoodOS brings the substrate it all runs on. The institution stays the institution. Nothing about running on it asks you to hand over what's already yours.
Deployed, not a diagram
I've sat through a lot of impressive systems that only existed in a pitch. A clean architecture diagram and a roadmap. Nothing you could actually run.
This isn't that. It's deployed. It's handling real data, at real scale, under the rules, right now. When I show you something in a demo, you're looking at a thing that's running — not a thing I'm hoping to build with your money.
If you want to see what it runs on, and why moving it stays a decision you get to keep making, that's a conversation worth having.