MIT-licensed · single Go binary

One dashboard for every Proxmox host you run

Ferrum rolls every cluster and standalone node into one live view: inventory, backups, HA, firewall, alerting, and an assistant that can act on it, instead of a browser tab per host.

Walk through the demo Install Ferrum

1binary, no runtime deps

SQLite / Postgresyour choice of store

12look-and-feel presets

Ferrum's fleet overview dashboard, showing node and guest counts, CPU and memory load, and active alerts across a cluster.

Fleet health

98.4%

Guests online

26 / 26

Self-hosted, your infrastructure only PVE clusters, standalone nodes & PBS remotes REST API & MCP for agents and scripts Built-in AI assistant, works fully offline

Everything a Proxmox fleet needs watched, in one place

Ferrum isn't a prettier login screen bolted onto the Proxmox API. It keeps its own inventory, its own alert history, and its own audit trail, so it stays useful even when a node is unreachable.

Fleet-wide overview

Every connection (cluster, standalone node, or PBS remote) rolled into one dashboard: node and guest counts, CPU/memory/storage, active alerts. Rearrange it from 20+ widgets and it stays that way per user.

Alerting & automation

Threshold alerts on CPU, memory, disk, and guest state; config drift detection; guest lifecycle policies; capacity forecasting; a fleet health score; scheduled health-digest emails; and Terraform/Ansible inventory export, so you find out from Ferrum, not from a user.

Inventory & operations

Nodes, VMs, and LXCs with live consoles, snapshots, and a guest agent file browser.

noVNC + xterm.js consoles
Bulk start / stop / migrate

Storage & backups

Pool usage and Ceph health across every connection, backup job status, replication, and PBS integration alongside native PVE storage.

HA, firewall & SDN

HA groups and resources, cluster and per-node firewall rules, and SDN zones/VNets/subnets, all scoped per connection.

Integrations

Outbound webhooks and a REST API and MCP server with scoped, per-user keys for third-party tools and agents.

Access & auditing

Per-user roles, optional OIDC single sign-on, session monitoring, and a full audit log of every mutating action, UI, API, and MCP alike.

AI Assistant

Chats with the same tool catalog MCP exposes, using any OpenAI-compatible provider, or Needle 2, a bundled local model with no API key or network needed.

The real thing, not a mockup

Screenshots below are Ferrum running against an actual fleet: three connections, six nodes, twenty-six VMs and LXCs across two clusters and a standalone host, with Ceph and NFS storage behind it.

Topology graph showing clusters, nodes, and guests as a connected map. — Topology, the fleet as a map, not a table

Inventory table listing nodes, VMs, and LXCs with live status. — Inventory across every node

Backup jobs and replication status. — Backup jobs & replication

High availability groups and resources. — HA groups & resources

Cluster firewall rules. — Cluster & per-node firewall

Twelve look-and-feel presets ship in Settings, each with light and dark and a choice of accent color; pick one and it applies everywhere, including consoles and the AI assistant.

EnterpriseProxmox-nativeTerminalGlass Flight DeckMidnightPaperGlassmorphismNeumorphismBrutalistSolarizedHigh ContrastAurora

See it before you install it

The live demo is a full click-through of the real interface, seeded with a fleet already running: no signup, no Proxmox host required.

Open the demo Read the docs