omacloud machine sizes and regions
two sizes, both running the same omarchy. the numbers for each, what comes with every machine, and what we can and can't tell you about regions yet.
two sizes on sale. the same omarchy on both.
- standard
- vcpu
- 2
- ram
- 4 gb
- disk
- 80 gb
- gpu
- none
- price
- $30/mo
- plus
- vcpu
- 4
- ram
- 8 gb
- disk
- 160 gb
- gpu
- none
- price
- $50/mo
each size is a virtual machine that belongs to one customer: nobody else logs into it, and its disk holds only your files. the vcpus are shared-cpu virtual cores rather than physical cores reserved for you. the disk is the machine's own and stays with it for as long as the machine exists.
those are the only two sizes we sell today. there are no larger sizes, no gpu sizes and no custom builds. what each size costs, and how to choose sits with the plans.
what you get whichever size you pick.
- a pinned, tested omarchy release — what omarchy is, in its makers' words.
- a streaming server on the machine, paired to moonlight on each of your devices with a four-digit pin.
- your own user with full sudo, and ssh with your own key. password login over ssh is off.
- a firewall open only to the stream and ssh. there's no web desktop and no web port.
- no gpu. the desktop is drawn, and the stream encoded, in software on the machine's vcpus.
- no automatic backups or snapshots today. keep what matters in git or synced elsewhere.
the full sequence, from payment to the first stream, is in how omarchy in the cloud works.
what the numbers mean for your work.
the stream runs on the same vcpus as everything else, because there's no gpu to hand it to. in our proof of concept, on standard's two vcpus, the encoder used about a quarter of the cpu, as measured on the home page. the rest is left for your own work.
memory is usually the first thing to run short. containers, a local database, a language server on a large codebase and a browser full of tabs each want their share, and plus has twice what standard has. disk is the other difference: plus has twice the space, for bigger repositories, images and build caches.
neither size is built for graphics-heavy work. without a gpu, 3d apps, games and video editing are drawn in software, and a bigger size won't change that.
you choose the region. once.
when you create a machine, you pick the region it lives in from the list the app offers. that choice is permanent: a machine can't be moved between regions, so moving means building a new one.
pick the region nearest to where you'll use it. a streamed desktop sends every keystroke to the machine and every frame back, so the distance between you and the data centre shows up in every keystroke. a bigger size can't buy that back.
we're not listing regions on this page yet, on purpose. which regions are open at launch is still being confirmed, and a list that names a city we can't yet serve would be a promise broken on the first day. when it's settled, it goes here. until then, the waitlist form asks where you are — and the answers help decide where we open next.