Proxmox alternative: Phillux compared

Two starting points, two builds

Anyone wanting to run virtual machines on their own hardware today comes across Proxmox VE sooner or later — an open-source project cut for the data centre: many nodes, distributed storage, a cluster layer as the foundation.

Phillux grew out of a different starting point: the single server that has to run a few machines reliably, and that may one day become two. Different decisions follow from that — fewer parts, fewer preconditions, fewer steps to the first running machine. Whether those are the right ones for you depends on what you have in mind; this page lays both out.

The three differences you can put your hands on

What actually stands out day to day when you come from Proxmox.

Images and ISOs in Phillux: Debian netinst, virtio-win drivers and a Windows 11 image

Windows 11 without firmware groundwork

Windows 11 demands UEFI, Secure Boot active and a TPM 2.0. In Phillux all three are there the moment a guest is created as Windows — there is no step in which you set firmware type, EFI disk and TPM one by one. The installation then runs unattended to a finished desktop with an administrator account, and the virtio drivers are already inserted.

A firewall that cannot touch the host

In Phillux the rules hang exclusively off the network device of the guest in question. The host's own ruleset is shown, but never changed. The practical effect: a wrong rule takes at worst a single guest off the network — your own access to the server survives either way. The same goes for the host's network cards, which Phillux only displays and never reconfigures.

Cluster view in Phillux with nodes, free capacity and a suitability score

The cluster is optional, not the foundation

Phillux runs as a single service. On one host there is no cluster stack that has to stand first, and no second database wanting maintenance — backing up here means copying one file. When a second host arrives, the two are joined with a single key; one node lays the configuration down and the others reconcile against it. With shared OCFS2 storage, a running VM live-migrates between the nodes.

Phillux and Proxmox VE side by side

Verifiable properties only — the rows where the two genuinely differ.

As of 2026. Statements about Proxmox VE refer to the freely available edition without a support subscription.
Topic	Phillux	Proxmox VE
Guests	Virtual machines and LXC containers	Virtual machines and LXC containers
Underpinning	Debian 13, a single service	Debian with a kernel of its own, several services
Windows 11 guest	UEFI, Secure Boot and TPM 2.0 automatically; unattended installation	Firmware, EFI disk and TPM are set one by one at creation
Firewall	Per guest only; the host ruleset is merely displayed	At data centre, node and guest level
Cluster	Optional; two nodes are enough, configuration reconciled from a leading node	A fixed part of it; quorum needs three nodes or a quorum device
Live migration	Over shared OCFS2 cluster storage	Over shared storage such as Ceph, ZFS replication, NFS or iSCSI
Storage	Several storages, separated by content type, with a default per type	Several storage types, among them Ceph and ZFS
Sign-in	The server's own accounts; Phillux stores no passwords	Its own user management, with PAM, LDAP and further sources
Interface	REST API, documentation generated from the program itself	REST API with documentation of its own
Licence and supply	Commercial, from 190 € per server and year; Enterprise per 16 cores	Open source, no licence cost; subscription for support
Support	Straight from the maker — from the people who wrote the code	Community forum; vendor support through the subscription

What Phillux is built for

So that nobody starts a trial that cannot fit: here is the cut, stated plainly — what Phillux is made for, and where its scope ends.

One to a handful of hosts. On a single server Phillux runs with no cluster stack at all; from two nodes upwards live migration comes with it. It is not laid out for double-digit node counts with high-availability groups and several sites.
Shared storage over OCFS2. That is what live migration rests on. A CephFS can be mounted as a storage; storage distributed across many nodes, and ZFS with replication and checksums, are not part of the scope.
Windows and Linux guests side by side. UEFI, Secure Boot and TPM 2.0 stand ready for Windows guests by themselves, and the installation runs unattended. This is where the cut saves the most time.
An operation meant to stay simple. One service, one file, one autostart — on a single host, backing up means copying that file.
Commercially licensed, per server and year. If you have to read the source, adapt it, or run it without a licence agreement, Phillux is not the right thing — and a trial will not change that.

If your setup stays inside that frame, you can do the sums in a month with real machines instead of data sheets.

Bringing existing guests across

A move does not mean setting everything up again. Point Phillux at a Proxmox or VMware ESXi source and it brings the guest across: disks, specs and network in one step. It is the same flow as creating a new virtual machine, only starting from an existing machine instead of a blank one.

For a first impression a single server is enough: a script sets Debian and Phillux up, autostart and an encrypted connection included. The trial period runs for a month, without registration and without functions being held back.

Try it fully for a month Tell us how many processor sockets your server has and what should run on it — we'll name the right edition and the price. For the trial we need no personal details.

Download it See the editions

Frequently asked questions about the comparison with Proxmox

|

Can I bring my VMs across from Proxmox?

Yes. Point Phillux at the Proxmox source and it brings the guest across — disks, specs and network in one step. It then keeps running under Phillux. The same path works for VMware ESXi.

Do I need three nodes for a cluster?

No, with Phillux two are enough. One node lays the configuration down and the others reconcile against it, and which node leads can be turned around at any time. Live-migrating a running machine additionally needs shared cluster storage (OCFS2) that every node can reach.

Does Phillux support Ceph or ZFS?

In part. A CephFS can be mounted as a storage — it stands in the storage dialogue next to a directory, NFS, SMB, iSCSI and OCFS2. What Phillux does not have is distributed Ceph storage across many nodes, or ZFS with replication and checksums; for live migration it relies on shared OCFS2 cluster storage. Anyone who needs storage distributed across many nodes will not find it in Phillux.

Is Phillux open source?

No, Phillux is a commercial product and is licensed per server and year. It runs entirely on your hardware: no cloud holding your machines, backups or credentials, and no telemetry. Only the periodic licence check goes out — and if that ever fails to reach us, Phillux carries on while a grace period runs.

What does Phillux cost in comparison?

Proxmox VE is given out as open source with no licence cost, and support runs through a subscription. Phillux is commercial and licensed per server and year: 190 € for Essentials, 390 € for Standard and 490 € per 16 cores for Enterprise, depending on the size of the installation — sockets, cores and the number of hosts in a cluster. Support is chosen separately and costs nothing unless a response time is wanted. Anyone who would rather compare first can try it fully for a month.

Does Phillux run on the same hardware as Proxmox?

As a rule, yes. What it needs is an x86-64 processor with the virtualization extensions enabled in the BIOS, and Debian 13. For new hardware there is an installation medium that sets up Debian and Phillux unattended.