Sign in Sign up

Platform

Virtual Private Servers

A brintOS machine runs in your browser tab — until you'd rather it didn't. Suspend it and resume it on a brintOS server, instantly, mid-process: the shell prompt, the half-finished build, the open TCP connections all carry over. Close the laptop and the machine keeps running server-side. Pull it back into a tab tomorrow, exactly where it is by then. One machine, two places to run it.

The tab is a great computer with one flaw

Running Linux in the browser means zero setup, your own hardware at full speed, and nothing to rent while you're at the keyboard. But a tab lives and dies with the tab: close the lid, lose the connection, or just navigate away, and the machine stops. That's the wrong deal for a six-hour build, an overnight test run, a service that should answer while you sleep, or a notebook kernel holding a day of state.

A brintOS Virtual Private Server is the same machine given a second home. It isn't a separate product you provision, image, and maintain — it's a place your existing machine can be. Suspend in the tab, resume on the server; suspend on the server, resume in a tab. The machine doesn't know the difference, and nothing running inside it restarts.

Suspend here, resume there

  • Mid-process, not re-boot. Suspension captures the whole live machine — kernel memory, every process, open file descriptors, device state. Resume is not a boot: your compiler is still on the same object file, your tmux session is still laid out, your psql transaction is still open. A handoff typically completes in a few seconds — most of it just the remaining memory pages crossing the wire.
  • It works because the machine is portable by construction. A brintOS machine is a wasm32 Linux kernel and wasm userspace against content-addressed storage. The same bytes execute identically in your browser and on our servers — there is no cross-architecture translation, no "cloud image" that drifts from the local one. A suspended machine is just data, and data moves.
  • Storage doesn't move at all. The root filesystem already lives in cloud storage; your tab and our servers are both just clients of it. A handoff transfers the machine's live memory and device state, not its disk — which is why it's seconds, not a backup-restore.
  • Both directions, any browser. Send a machine server-side from your desktop and pull it into a tab on your laptop later. Detach and reattach from anywhere; the machine's home is your account, not any one device.

What changes when it runs server-side

Nothing, from the inside — and a few useful things from the outside. A server-resident machine keeps its identity: same snapshot lineage, same org VPN address and internal DNS name, same collaborator permissions. Teammates' machines reach it exactly as before. You can watch or drive it through Remote Desktop from any tab, as can anyone you grant. And because it no longer depends on a browser being open, it's a fit for the always-on jobs a tab can't hold:

  • Long builds and test runs. Kick off the job locally, hand the machine to the server, close the laptop. Reattach over coffee to a finished build — same terminal scrollback and all.
  • Always-on services. The team package mirror, staging database, or bot that must answer at 3 a.m. runs server-side on the org network, launched and administered like any other machine — because it is one.
  • Scheduled work. Wake a suspended machine on a schedule, let it run its cron jobs, and suspend it again — paying for minutes of compute instead of a month of idle.
  • Handoffs between people. Park a machine server-side at the end of your day and a colleague in another timezone pulls it into their tab — live state, not a snapshot-and-relaunch.

Where it runs and what it costs

Server-side machines run on the same regional infrastructure as the NIC exit nodes, so a machine usually resumes in the region nearest wherever it was suspended. Tab-side execution stays free on every plan — it's your hardware. Server-side time is metered by the vCPU-hour and RAM-hour while the machine is actually resumed; suspended machines cost only their storage, which rounds to nothing. Team plans add org-owned server machines with pooled billing, and Enterprise runs the whole resume fleet inside your own VPC.

The point

Local-first computing usually means choosing: the immediacy of your own hardware, or the permanence of a server. A machine that suspends in one and resumes in the other, live, means you stop choosing. Work at the speed of your own CPU all day, then let the machine outlive the tab.