Proxmox infrastructure and server hardware management in one dashboard.
Depl0y puts every Proxmox cluster and standalone host you run into a single view, alongside the iDRAC and iLO data for the machines underneath them. It is a self-hosted panel that sits next to the Proxmox web interface rather than replacing it.
Quick start · Screenshots · Watch walkthrough · Documentation · Releases · MSPReboot
One dashboard over three registered sites — two clusters and a standalone lab host.
▶ Watch the 3½-minute walkthrough
· 55-second highlight
· captions
A real screen recording against the isolated demo instance — captions on screen, no narration audio.
One view over every Proxmox environment. The Proxmox web interface is scoped to the cluster you log into. Depl0y registers each cluster or standalone host as its own endpoint, with its own credentials, and puts the totals, node health and guest inventory for all of them on one page.
The hardware, not just the hypervisor. Depl0y polls Dell iDRAC and HPE iLO over Redfish and shows health, power draw, temperature, DIMM and drive inventory and firmware versions next to the VMs running on that machine — including the BMC power path, the only one that can turn a machine back on once it is off.
A narrower door into Proxmox. Depl0y has its own accounts with Admin / Operator / Viewer roles, TOTP and an audit log, so operators get a login here instead of Proxmox root credentials. Depl0y itself still connects with a privileged token; it narrows who touches the hypervisor, not what Depl0y can do.
Know what the hardware is doing. Redfish health, draw and temperature for every BMC-equipped server, then one machine in full detail.
Follow performance down to a single guest. Live gauges and an hour of history, drawn from the node's own RRD data.
See how the estate fits together. Every site, node, guest, bridge and BMC as one graph.
Compare environments side by side.
Then get the day-to-day work done. Find any guest across every site, and place a new one deliberately.
Keep a guest running when its node fails. The failover wizard checks where each guest's disks live and whether the cluster can fence a dead node, then applies one fixed, best-practice HA setup and shows exactly what it will change.
📸 Browse the full gallery — 32 views, light and dark
Depl0y talks to the standard Proxmox API. It is an additional console, not a replacement — the two are meant to be used together.
| Proxmox web interface | Depl0y | |
|---|---|---|
| Guest lifecycle, config, snapshots, console on one cluster | Yes — the authoritative tool | Yes |
| Several clusters or sites in one view | One cluster per login | All registered endpoints on one page |
| iDRAC / iLO health, power draw, temperature, firmware inventory | Not part of the Proxmox UI | Built in, over Redfish |
| Powering on a host that is fully off | Reach the BMC separately | BMC power actions in the same UI |
| Making a guest fail over to another node | Set up HA resources, rules or groups and replication yourself | Failover wizard: checks storage, quorum and passthrough, then applies a fixed setup |
| Cluster creation, storage configuration, PVE upgrades, repositories | Yes — do it here | Not covered; use Proxmox |
| Accounts for the panel itself | PVE realms and ACLs | Own Admin / Operator / Viewer roles, TOTP, audit log |
If you only run one cluster and never touch the BMCs, the Proxmox web interface is likely all you need.
On a small Ubuntu or Debian VM of its own — not on a Proxmox node:
curl -fsSL https://raw.githubusercontent.com/agit8or1/Depl0y/main/install.sh -o install.sh
less install.sh # read it before running it
sudo bash install.shThe installer sets up Python, Node.js, nginx and SQLite, generates SECRET_KEY
and ENCRYPTION_KEY into /etc/depl0y/config.env, installs the
depl0y-backend systemd unit and configures nginx. It downloads the latest
GitHub release (depl0y.tar.gz) and
checks it against the release's minisign-signed SHA256SUMS (see
Verifying releases), so github.com must be reachable.
Then open http://<server-ip>/ and sign in with admin / admin.
Change that password immediately and enable TOTP.
Would rather not run it yourself? MSPReboot offers managed hosting and commercial support for Depl0y — see mspreboot.com. Self-hosting stays free and MIT licensed either way.
Create an API token in Proxmox (Datacenter → Permissions → API Tokens) with
privilege separation unchecked, then in Depl0y go to Proxmox Hosts → Add
Datacenter and enter the hostname, port 8006, the token's owner (e.g.
root@pam), the token ID and its secret. Leave Verify SSL off for the default
self-signed Proxmox certificate. Repeat for every cluster you want in the
dashboard.
Full walkthrough, including BMC setup: docs/QUICKSTART.md.
| Depl0y host | Ubuntu or Debian, 2 vCPU / 2 GB RAM / 20 GB disk |
| Network | 8006 to the Proxmox API, 443 to each BMC, 22 to nodes for import and terminal features |
| Outbound | github.com for in-app updates, downloads.dell.com for the optional Dell firmware check |
| Browser | outbound access to cdn.jsdelivr.net for the noVNC console and to OpenStreetMap for the map view |
- Failover wizard — make a VM or container restart on another node if its node fails. Shared storage or ZFS replication is chosen automatically, guests on node-local storage are explained (with a disk move offered where possible), and the protection can be removed again from the same place.
- Cloud images — Ubuntu 20.04/22.04/24.04 LTS, Debian 11/12 and Rocky Linux 8/9, with cloud-init for hostname, user, SSH key, static IP and packages.
- VM import — upload OVA, OVF, VMDK, VHD, VHDX, QCOW2 or RAW, or pull VMs
straight from an ESXi host or vCenter; disks are converted with
qemu-img. - LLM deployment — run an open-source model as a VM. Pick from a catalogue of open-weight models (Llama 3.2, Qwen 2.5, Gemma 2, Phi-3.5 and others) and Depl0y builds the guest, installs the serving stack — Ollama, llama.cpp, vLLM or LocalAI — and sizes it for the model, with optional NVIDIA or AMD GPU passthrough.
Details in the feature reference.
Depl0y uses the standard Proxmox VE API v2 (/api2/json) and does not gate on a
Proxmox version string. It is developed and exercised against Proxmox VE 8.x;
other releases are not routinely tested. The failover wizard supports both
Proxmox VE 9 (HA rules) and 8 (HA groups). BMC support covers Dell iDRAC 7, 8
and 9 and HPE iLO through Redfish v1, with an SSH fallback for hosts whose
BMC does not answer Redfish.
- The installer supports Ubuntu and Debian only. A
docker-compose.ymlexists in the tree but the installer is the supported path. - Depl0y is exercised with a Proxmox token that has privilege separation
disabled, so it carries the owner's full rights.
pvesh-backed actions (OS-level power, node terminal, cluster join) effectively needroot@pam. A least-privilege Proxmox role has not been validated. - VM import, inter-node SSH setup and the node terminal need SSH access to the Proxmox nodes.
- The firmware version check is Dell-only — HPE firmware is inventoried but not compared against a catalog.
- Depl0y stores its own state in SQLite on a single host; there is no clustering or HA for the panel itself.
- The noVNC console and the map view need outbound browser access, so the panel is not fully air-gapped.
- Proxmox passwords, API token secrets and BMC passwords are stored
Fernet-encrypted; the key lives in
/etc/depl0y/config.env. - Accounts support TOTP with backup codes, and every action is written to the audit log.
- Report vulnerabilities privately — see SECURITY.md. Please do not open a public issue for a security problem.
Upgrade in the panel: Settings → System Updates → Check for updates →
Install, which reads the latest GitHub release. Re-running install.sh
does the same and preserves the existing ENCRYPTION_KEY.
Back up before upgrading:
sudo systemctl stop depl0y-backend
sudo tar czf depl0y-backup-$(date +%F).tar.gz \
/etc/depl0y/config.env \
/var/lib/depl0y/db/depl0y.db
sudo systemctl start depl0y-backendconfig.env holds ENCRYPTION_KEY. Without it the stored Proxmox and BMC
credentials cannot be decrypted, so back up both files together.
- Questions and bugs — GitHub Issues (there is also a Report Bug form in the panel).
- Contributing — CONTRIBUTING.md. For anything large, open an issue first.
| Quick start | install, requirements, connecting Proxmox and BMCs, troubleshooting |
| Feature reference | the full catalogue |
| Deploy LLM guide | running an open-source model as a VM, end to end |
| Screenshot gallery | 27 views, light and dark |
| Architecture | internals, background jobs, state on disk, running from source |
| Proxmox API tokens | token creation in detail |
| Capture tooling | regenerate the screenshots and video |
| Changelog | release history |
Depl0y is built and maintained alongside other operations tooling at mspreboot.com, an MSP consulting practice focused on operations, profitability and growth.
Depl0y is MIT licensed and free to self-host — nothing here requires an engagement, and the project stays open source.
If you would rather not run it yourself, MSPReboot offers managed hosting and commercial support for Depl0y. Get in touch through mspreboot.com for scope and pricing.
MIT licensed — see LICENSE.
Not affiliated with or endorsed by Proxmox Server Solutions GmbH, Dell or HPE.
🐺 Supervised by Phil the Husky.







