Run your own cloud.
Turn the Proxmox and OpenStack clusters you own into a cloud you sell, under your brand and on your billing.
Proxmox VE · OpenStack · platform v2.80.11
Three capabilities, one control plane.
Each capability is designed for a different kind of organization, and all three run on the same platform.
Unified management
For hosting providers and managed service providers. Clusters, customers and invoicing in a single platform, under your brand.
- Management of every Proxmox VE and OpenStack cluster from one dashboard, across all sites.
- Customers, servers, networks and storage administered from a single interface.
- Your branding on the storefront, customer dashboard, e-mails and invoices.
- Monthly invoicing through your own Stripe account, with EU VAT handled.
GPU acceleration and provisioning automation
For GPU and neo-cloud providers. Customers order GPU servers on demand; the platform provisions and bills them automatically.
- Servers with one to eight GPUs, on your own hardware.
- Customers select a configuration and see the monthly price before ordering.
- Your price list shows cost, price and margin for every item.
- Customers can manage servers through their own AI assistant, using the MCP endpoint.
Policy engine and organization management
For regulated sciences. Control who may use which data, retain evidence of every decision, and give each lab its own workspace.
- Institutions, labs, projects and teams, each with its own members and roles.
- Data access is granted separately from compute access.
- Policy templates for GDI, ELIXIR and EU medical research.
- Policies can be trialled before they are locked, with a scheduled effective date.
- Every decision and administrative action is recorded in a tamper-evident log.
Unified monitoring and observability.
One view of health, alerts and metrics across every site, from the hypervisor and its BMC up to each customer’s server.
Monitoring stack per cluster
Prometheus, Alertmanager and Grafana run on each Proxmox VE cluster’s own management network, deployed from code. Data stays on site, and alerts keep flowing when the control plane is unreachable.
Preloaded dashboards
Grafana comes with dashboards for Proxmox, hypervisor hosts, network probes, BMC sensors and GPUs, for immediate visibility from day one. Metrics are collected every 30 seconds and retained for 90 days by default.
Preconfigured alerting
The stack ships with 23 alert rules for nodes, quorum, storage, backups, EVPN peers, BMC sensors, certificates and GPUs. Alerts are delivered through your own mail relay.
Service health
One Monitoring page shows stack health, firing alerts and unreachable targets for every cluster, with direct links into Grafana and Alertmanager.
Maintenance-aware alerting
Nodes marked as planned offline are excluded from every down alert, so planned work raises no false alarms.
Server metrics
CPU, memory, network and disk for every server, from the last hour to the last 30 days, for your team and your customers. Read from the hypervisor on site; the control plane stores none of it.
Usage telemetry
Metered usage is written to an append-only ledger that feeds billing directly and survives outages.
Open, standards-based endpoints
Standard Prometheus endpoints let you export the same metrics to the observability platform you already run.
One foundation under every cloud you run.
Your hardware stays yours. The platform reaches it through one appliance per site, and nothing on your network has to accept a connection from outside.
Clusters
Connect running Proxmox VE and OpenStack clusters without rebuilding them, or start a new site from a clean host.
Site appliance
Dials out over a private mesh. No inbound ports, and cluster credentials stay on site.
Isolation
Each customer account gets its own VPC, separated in the kernel with EVPN. Servers start private.
Consoles
VNC, serial and an SSH terminal in the browser, on a certificate that expires in five minutes.
Storage
Snapshots, data disks by storage class, and S3-compatible buckets with a file browser.
API
Scoped API keys with expiry, a REST API, and an MCP endpoint on your own host.
Updates
Site appliances update on canary and stable rings, with rollback and a guard against downgrades.
Backups
Platform state is backed up every night, pulled off-site, and restore-tested.
Choose where the control plane runs.
Hosted and on-premise run the same platform, priced per node per month. For large deployments, universities and research institutions, contact sales.
Hosted
€490per node / month
Explode-hosted control plane. We take care of upgrades, backups and maintenance.
- Control plane
- Run by Explode
- Clusters connect
- Site appliance, outbound only
- Updates
- Managed by Explode
- Customer billing
- Stripe, EU VAT
On-premise
€690per node / month
Self-hosted control plane on your own infrastructure, licensed offline with no call home.
- Control plane
- On your own VM
- Clusters connect
- Directly, on your network
- Updates
- Signed bundles, one click
- Customer billing
- Optional
Custom Deployments
Contact sales
Multi-node deployments, universities and research institutions.
- Deployment
- Hosted or on-premise
- Scale
- Multi-node, multi-site
- Education and research
- Education licence
- Pricing
- Quoted per contract
Prices exclude VAT. A node is one hypervisor host.
See it running on your hardware.
A demo is a working cloud, not a slide deck: your hardware profile, your brand, a server created and an invoice with your name on it.