Operations layer for Proxmox

AtlasPVE is the safe operations layer for Proxmox.

VMs, storage, networking, updates, backups and self-hosted services are brought into one visual interface. In critical flows, impact, confirmation and rollback information are visible before changes are applied.

Local installation

No blind operations

Impact visibility in critical flows

Thanks to the Proxmox community

Over the past month, visitors from all over the world looked at Atlas, tried it and wrote in. The questions, the bug reports and the suggestions carried the product to where it is today.

and many more

FEATURE · ANATOMY

The structure beneath one screen.

Atlas is built around a clear operations flow from hardware to interface. Data is read bottom-up; decision and execution steps are followed on the same screen.

data flows bottom to top

Proxmox-native

Open-source agent (AGPL)

Independent product

Where to start?

Product

Atlas product map.

Virtual machines, storage, updates and safe operations flows meet in the same local management layer. The product installs on the Proxmox host; core capabilities are gathered in one interface.

Proxmox host

Proxmox VE

Atlas

operations layer

The difference

Before Atlas / With Atlas

Same infrastructure, two very different experiences.

Without Atlas

With Atlas

Dense Proxmox tables

Risky CLI copy-paste

Opening ports on the router

Hard to see storage/network relationships

No easy app experience

Hours of manual management, time and effort spent

Long guides, trial and error, risky critical operations

Hidden processes, invisible operations

apt upgrade without knowing what each package touches

Unexpected reboots, silent service interruptions

Risk of not booting after a kernel update

Continuous manual health checks are required

Open SSH and forward a port just to get into a VM

Visual topology

Guided storage/network operations

Operation preview and rollback awareness

Controlled service management from the local interface

Atlas Apps catalog experience

Operational load decreases; time stays with the team

Warnings and protection in critical flows; best practice up front

No invisible operation, no hidden process, no hidden service

Every update’s dependencies, effects and changelog, up front

Reboot/service impact grouped in advance; guided with a snapshot suggestion at the risky step

Post-kernel boot guard: if boot isn’t confirmed, it auto-reverts to the last working kernel

Atlas Watch delivers daily summaries and critical alerts automatically

In-VM terminal, file browser and metrics, with no SSH, no open port

Operational depth beneath the visual interface.

Every screen does real work, nothing decorative.

Visual Topology

The resource chain of the server is drawn end to end: from the machine to its core and memory share, from the disk to the storage, from the pool to the physical drive. The flow along the lines is live traffic. Separate maps exist for storage, network, processor, containers, access and updates.

Safer Storage

ZFS, LVM and Btrfs in one visual language. Recover an accidentally destroyed pool, a stopped RAID array or a disconnected store straight from the on-disk label, without writing data. Orphan-store and capacity-trend warnings stay up front.

Safety Control for Network Changes

Bridge, bond, VLAN and OVS changes are prepared on a visual map. After applying, a confirmation window starts; if connectivity is not confirmed, the configuration is rolled back automatically.

Safer Updates

Every package is tagged by impact (seamless / service / reboot / kernel) and installed in ordered rounds. Pre-install snapshot, pinning the running kernel and boot verification: if a new kernel won’t boot, the system comes back on the old one.

Local Access

The Atlas management surface runs on the Proxmox host. The product can be used without an external control panel or a permanent internet connection.

Atlas Apps

An application installed from the catalog lands with its folders and published ports in plain view. A custom compose file can be added as well. Before installation, boundary crossing settings such as privileged mode, mounting the Docker socket and using the host network are flagged. The application can also take its own address on the network and be reached by name.

Hardware passthrough

PCIe and USB assignments are handled through a guided interface, with full support targeted even for cards NVIDIA restricts from vGPU. If something breaks, a support report can be generated with sensitive data removed.

More than a task manager

Inspired by the Windows Task Manager experience, even non-Atlas processes become visible and manageable: pause, reprioritize or terminate. System-critical processes stay protected until unlocked with a password; action explanations and change records stay visible.

SSH-free VM access

Terminal, file browser and detailed metrics inside the VM, with no SSH and no open network port. Atlas uses the QEMU Guest Agent host-guest channel: no new attack surface, no need to memorize the VM IP, and modules are added through a controlled installation flow.

Smart recommendations

Atlas regularly evaluates Proxmox best-practice rules: is the protection lock off, are snapshots piling up, is the disk cache safe, does the CPU type block live migration... It turns 50+ rules into prioritized, actionable recommendations for each VM, processor and update.

Protection Layer

In critical flows, warnings, confirmations, audit records or rollback points are activated. Expert users can proceed deliberately in flows that allow it after seeing the warnings.

Machine Setup and Templates

A machine built from a cloud image comes ready on first boot: the user, SSH key, network name and selected packages are placed during setup, with no ISO wizard to sit through. Templates are managed in one place. Ready workload profiles, from media server to database to game server, bring core and memory suggestions at setup time.

Access and Permissions

Who reaches which resource with which role is visible on a single map. Proxmox roles are used, custom roles can be defined where needed, and permissions are assigned per path. Two step verification and API tokens are managed from the same screen. Atlas builds no user system of its own, it reads permissions from Proxmox.

And more

Processor

Down to the core.

See where each VM's virtual cores are actually running right now, across physical cores and threads, read one-to-one from the host's /proc, not a Proxmox estimate. Flowing lights carry the real load; a sudden spike shows up as a single spark.

Watchdog

Atlas Watch monitors infrastructure health regularly.

Atlas Watch gathers server health into a single summary email at the configured time: pending updates, disk and memory pressure, VM problems, reboot requirements. If a critical finding appears, it sends an alert without waiting for the summary.

Atlas Watch

watchdog

Mail

local path

Admin

Team

On-call

Daily summary

Critical · instant

See the impact before applying changes.

Atlas makes risk, impact and confirmation steps visible in flows that touch infrastructure.

Impact and Confirmation

In critical flows, planned steps, affected resources and rollback options are shown before changes are applied to Proxmox.

Local Agent

Atlas runs on the Proxmox host as an open-source local agent, nothing connects in from outside.

Local Runtime

The product runs on the Proxmox host; management data stays inside the customer infrastructure.

Support Report

Failed operations can generate a support report with secrets stripped out.

Atlas runs on the local Proxmox host.

Atlas is not an ordinary control panel. It is a local operations layer that touches storage, network, update, backup and VM operations on Proxmox. The product runs inside the customer’s own infrastructure; data does not leave it.

The engine that runs with root privileges (the agent) is open source, AGPL. What runs on the server does not have to be accepted blindly: the code can be inspected and its behavior audited.

Every release ships signed. Security rules, tested agent behavior, rollback logic and version-specific protections are kept fresh through updates.

Open source, but where is the source?

The part with the most power over the system is the agent, because it runs as root on the host. With the install it settles into the host as readable source, so the code can be opened on the machine it runs on and read line by line. Nothing in the privileged layer has to be trusted as a closed file. It is not published as a public repository, and the reason is not only commercial. Atlas runs inside other people's infrastructure. Publishing every line of the privileged layer also publishes a map for anyone looking for a way in, and in an age where scanning is automated and constant, that map is read by machines long before it is read by people. The risk lands on the customer, not on the vendor. Reading the code on the machine where it runs gives the same assurance without handing that map to everyone at once. Whoever installs Atlas can audit exactly the code that is executing on their own host, which is the copy that actually matters. Responsibility does not move either way. A hole in Atlas is ours, and closing it is ours.

The project page on GitHub

Frequently asked

Approach

We’re not robots, we’re human. Atlas’s interfaces are designed to make information easy to internalize, drawing on human psychology and pedagogy: what is happening on the server and the overall picture become genuinely clear. There’s no invisible operation, no hidden process, no hidden service.

We have deep respect for Unraid, ESXi, Synology, Docker and more. Atlas combines what those experiences teach with the virtualization power of Proxmox.

Moving from VMware to Proxmox: the guide

Pricing

Atlas makes local Proxmox infrastructure visible, auditable and safer to operate.

This is a purchase, not a rental: the paid version runs indefinitely. The update license is tied to its server; it can move to a new server, but not back to a previous one. Atlas keeps running on the old server but receives no updates. Keeping one more server up to date at the same time requires +1 update license. Annual renewal covers new features, security patches and Proxmox compatibility; if renewal stops, the installed version keeps running while new versions and update access stop.

Price lock

Only Home is on sale right now: $49/yr, locked at this price for life. Pro and Business are planned. First Home members get a discount when upgrading to a higher plan.

Explore Atlas without touching infrastructure.

The demo environment does not connect to a real Proxmox host. The Atlas interface and flows can be tried in a safe sandbox.

Open the Demo

Get in touch

Once the payment completes, the license and install link are ready in the account; installation is a single command.

Full name (optional)

Email address (for a reply, optional)

Message

Send

Sent

Thanks. The message has been received; if an email address was provided, we will follow up.

Could not be sent. Another attempt is possible shortly.

Make Proxmox operations visible and auditable.

First buyers are locking in their price.

Is Atlas open source?
The agent that runs with root privileges on the server is open under AGPL; what it does can be read and audited. The interface and product layer are offered under a commercial license, a deliberate choice to keep Atlas developed. The commercial product layer never runs as root.
Can Atlas run offline, without an internet connection?
Atlas runs offline, without an internet connection, and continues to operate locally on the Proxmox host. Internet is used only to fetch updates and the catalog; without it, the installed version keeps working fully.
Does Proxmox data pass through Atlas servers?
No. The Atlas management layer runs inside customer infrastructure; VM, storage and network content is not sent to Atlas servers.
What happens if I don’t renew updates?
Nothing gets locked. The installed local version keeps working; only new features, security patches and compatibility updates for new Proxmox releases stop. When renewal is enabled again, update access continues from that point.
Who did what, is it auditable?
Critical or privilege-raising actions (terminating a process, changing priority, destructive storage/network operations) are written to an audit log: who, what, when, result. The record is kept in two places: the system journald (immutable to non-root) and a separate file. Passwords and secrets are never logged.
Do my Proxmox users and roles still apply?
Yes. Atlas doesn’t build a parallel permission system; it mirrors Proxmox’s own user, role, group and API-token model. Proxmox decides who can see and do what, and Atlas follows it.
Is this a subscription, or am I buying the license?
This is a purchase, not a rental. The paid version runs indefinitely; without renewal, that version keeps running. Annual renewal covers new versions and update access.
What is a Proxmox operations layer?
Proxmox provides virtualisation; an operations layer takes on the work of running it every day. It shows in advance what a change will touch, makes the dependencies that run from storage to virtual machine visible, and turns a risky step into a deliberate one. Atlas does this locally: the management layer sits on the host itself and the data does not leave it.