Self-Managed VPS.
Your server. Your rules. Our iron underneath.

A VPS (virtual private server) is your own private slice of a physical server, with dedicated CPU, RAM, and disk, and full control of its operating system. A hypervisor divides one powerful machine into several isolated servers, so you get the control of a dedicated server at a fraction of its cost.

The server itself on non-oversold enterprise hardware, full root access, your choice of OS installed clean, one IPv4, the network firewall with IDS at the edge, DDoS protection, free daily snapshots, and the 500% SLA on infrastructure uptime. Not included: proactive monitoring, centralized logs, and hands-on management, which live in the Fully Managed addon.

Everything inside the OS: updates and patching, in-server firewall rules, application security, backup testing beyond the included snapshots, and fixing what breaks at the software level. If that list reads like freedom, this page is for you; if it reads like a burden, the managed page is one click away.

Yes, both, anytime, from the cart at order time or from your client area afterward. Fully Managed is $20/mo ($200/yr) and Dual Backups is 20% of your base server. Same server, no migration, no tier change; they are checkboxes, not commitments.

With an unmanaged VPS you rent the hardware and do everything else yourself: patching, firewall rules, backups, mail deliverability, and the 2 AM incidents. With a managed VPS the provider does that work for you. Unmanaged costs less per month; the difference is paid in your hours. Here you can start on either side and switch with a checkbox.

The two you will meet most are containers (LXC) and full virtualization (KVM). LXC containers share the host kernel: light and fast, but less isolated, and they cannot run their own kernel or Windows. KVM gives every server its own virtual hardware and kernel: true isolation, any operating system, and resources that behave like a real machine. Container-based plans are also the easiest to oversell.

KVM, exclusively. Every server is a full virtual machine with hardware-level isolation: your own kernel, any OS including Windows Server, and resources that exist physically, because we never sell the same resource twice.

Proxmox VE, the open-source enterprise virtualization platform, running KVM virtual machines. It is the same stack behind the free daily snapshots and the optional Proxmox Backup Server integration: battle-tested, auditable, and free of per-VM licensing costs that other stacks pass on to you.

Ubuntu, Debian, AlmaLinux, CloudLinux, and Windows Server 2025, installed clean at provisioning. After first boot the OS is yours: you update it, you configure it, you own it.

Any panel you like: install RemarkablePanel, cPanel, Plesk, DirectAdmin, Webuzo, or any of the open-source panels from our 49-app marketplace. Where a commercial license applies it bills at its standard price.

Yes. Every server is a full KVM virtual machine with root access and its own kernel, so Docker, Podman, and any container tooling run natively. Panels like Dokploy and Coolify install in one click from the marketplace.

Both run +3.00 GHz cores on the same hardware. Shared CPU gives excellent value for workloads with variable traffic. Dedicated CPU reserves every core for you, for checkouts, databases, and sustained loads that can never slow down.

They are a strong first layer, but they live on the same platform as your server. The Dual Backups addon (20% of your base) adds offsite PBS copies, restorable from the client area, for the failure modes snapshots cannot cover. Honest answer: production data deserves the addon.

Resizes are available on request from the client area, within your family or from Shared CPU to Dedicated CPU. Your data stays in place; the resources grow around it.

You can cancel anytime from your client area. Monthly billing means there is no lock-in, every plan carries a 30-day money-back guarantee, and you can take your data with you before the term ends. Hosting should be a service you choose to stay with, not one you are stuck in.

On enterprise dedicated infrastructure in US data centers, 100% NVMe and never oversold. The network in front of them carries the DDoS protection and the firewall with IDS included on every server.

The same as everywhere at RemarkableCloud: one hour of infrastructure downtime credits five hours back, automatically, from the first minute. It applies to the iron and the network regardless of who manages the OS.