ProxDash is a dashboard for a Proxmox fleet that fits on a phone. Connect it straight to a Proxmox VE cluster or node, a Backup Server, or a Datacenter Manager — in any mix — and to Portainer for the Docker guests inside them.
The iOS app is free to download, and viewing your fleet is free. Control actions on the phone are ProxDash Pro, an in-app purchase: $4.99 a year or $9.99 once.
Proxmox's own interfaces were built for a desktop. ProxDash puts your whole fleet — PVE, PBS, and PDM alike — on a screen you can actually read at the kitchen table, and adds the things they leave out.
PDM shows you a node is up. ProxDash shows you the SSD under it is at 91 percent life, the pool scrubbed clean on Sunday, and which service died at 3 a.m.
PDM passes through status and graphs. Through one read-only token per remote, ProxDash also reads each node's disks with their SMART verdict and remaining life, ZFS pool health and scrub results, the services that should be running, node certificate expiry, HA state and quorum, backup and replication schedules, and the journal.
An Overview of cards you drag, hide, and configure in place — with separate layouts for phones and wide screens, a Compact mode, and a getting-started card that walks a fresh install through setup.
No account, no analytics, no developer in the path. Requests go from you to your servers. Self-signed certificates get pinned by fingerprint after you approve them, and there's no way to switch verification off.
Sign in with a PDM account and ProxDash has PDM mint an API token under it — Auditor for view-only, Administrator for control. A token can never do more than its user, and your password is used once and never stored. The per-remote tokens that read disks and schedules are ones the server mints itself, read-only auditor tokens that can look and nothing else.
Sixteen rules. Unreachable remotes, nodes offline, storage past a level or forecast to fill, failed or stale backups, unhealthy containers, and newer images — plus the ones admins lose sleep over: a disk failing SMART or nearly worn out, a pool that isn't online, a service down, a node certificate expiring, HA in error or quorum lost, a failed replication or verification, and a reboot left waiting. Snooze an alert or mute one target, over ntfy, Telegram, or push, with a history of what cleared and how long it lasted.
The iPhone app, as it appears on the App Store.



Fleet, storage, backups, updates, and containers — read at a glance, controlled on purpose.
Fleet totals, top lists, storage, backups, updates, alerts, and Ceph health and capacity — drag to reorder, hide what you don't need, and configure each card in place.
Two weeks of PDM's usage history become bytes-per-day and the date each volume fills — so the warning comes before the disk does.
Per PBS datastore and per PVE backup job: when the last backup succeeded and whether recent ones failed, from a week of task history.
Guests move off, running backup jobs finish, the update runs over Proxmox's own terminal, the node reboots, and the guests come home — you watch every step. A replicated VM goes only where its copy lives, and a lone node reboots with its guests in place.
Each disk's SMART verdict and remaining life, ZFS pool health and the last scrub, the services that should be running, node and PDM certificate expiry, HA and quorum, and the journal when you ask for it.
Move a guest to another remote through PDM, choosing its storage and network bridge on the far side. From the web app today; on the phone in the next iOS update.
Link Docker guests to their Portainer environment: see stacks and containers, start/stop/restart them, and get honest image-update checks — digest-compared, pinned tags left alone.
Stop a stack cleanly before its guest, snapshot a guest, clean image layers — chained into scenes you can reverse, run on demand, or run on a weekday schedule.
ntfy, Telegram, and web push — all sent by your own server, with push keys it generates on first boot. Nobody else is in the path.
Every guest action, container action, and scene run lands in one history — including scheduled runs — and every task and container has a Log button.
The iOS app talks to your servers directly — no proxy needed — with a Face ID lock, a Home Screen widget, card layouts that sync through your own iCloud, and push alerts sent by your own server. Viewing is free; control is ProxDash Pro.
A Live Activity follows every node update and guest move in the Dynamic Island and on the Lock Screen. An Apple Watch app and complication show the fleet at a glance, Siri and Shortcuts check fleet status and run scenes, and the server's alert rules can be edited from the phone.
The server ships behind Caddy, which terminates TLS and is the only thing listening — the API publishes no ports and can't be reached in plaintext at all. Five wrong passwords lock the account for fifteen minutes, tokens are encrypted at rest with a key derived from your secret, and certificate verification can't be turned off.
Put it behind Tailscale or a VPN and nothing needs to be exposed. Proxmox, dashboard — the name is exactly what it says.
You need Docker and a Proxmox server to point it at — PVE, PBS, or PDM. The stack is pasted into Portainer's editor — no compose file lives on the host.
ProxDash plus Caddy, with a secret key and the one account's username and password.
The single account is created on first boot, and only while no user exists.
Settings checks each certificate, shows its fingerprint for you to confirm, then mints a scoped token under your account.
Caddy issues a certificate from its own CA by default — your phone trusts it after a one-time step Settings walks through — or gets a publicly trusted one with a DNS provider token. Migrations run themselves into one SQLite file under /data.