From single-node setups to multi-hypervisor clusters, the whole surface ships with every plan.
The full VM lifecycle on KVM and Proxmox VE, with the safety nets customers expect from a public cloud.
Deep dive into Compute & Instances →Deploy, suspend, resume, reinstall, migrate and destroy. Real-time VNC console, firewall rules and disk management.
Assign physical NVIDIA GPUs directly to instances for AI, ML and rendering. Bare-metal performance, billed hourly.
Live migration between hypervisors, HA monitoring with automatic evacuation, IPMI fencing and zero-downtime maintenance.
Point the panel at an existing PVE 8 or 9 node or cluster with one API token. Agentless over the Proxmox REST API, managed beside your KVM fleet with the same plans and billing.
Hyper-V enlightenments for KVM Windows guests, native Proxmox Windows support, and one-click repair of VMs imported from VMware or Hyper-V.
Adopt existing KVM domains as managed instances, or migrate whole fleets off Virtualizor with disks landing natively on LVM and ZFS.
Firecracker microVMs that boot in seconds and bill by the second: AI agent sandboxes, CI runners, and disposable compute your customers can sell today.
Deep dive into MicroVMs →Build from a Dockerfile, import a container image from any registry, or point at a git repo. Images are versioned and run on signed base images: Debian 13, Ubuntu 24.04/26.04, AlmaLinux 10, Rocky Linux 10 and Amazon Linux 2023.
A MicroVM gets a public IP, a VPC address, or both, exactly like a KVM instance. VPC-only MicroVMs leave through the VPC NAT gateway, with the same security groups.
Cap the lifetime at up to 8 hours, pause automatically on idle or mark always-on, and resume a paused MicroVM with its memory and disk exactly where it left off.
Account SSH keys are injected before first boot. A browser terminal, a streaming console log, live metrics, and HTTPS ingress on a wildcard sandbox domain or a custom domain.
The same wire protocol as E2B: existing Python and JavaScript SDK code runs against your own panel with two environment variables. Quick starts for Python, JS, the CLI, Anthropic and OpenAI.
GitHub App and GitLab runner connectors boot a fresh MicroVM per CI job and tear it down when the job ends. No idle runner fleet, no shared state between jobs.
The full public-cloud network toolbox, on your racks.
Deep dive into Networking & Load Balancers →Isolated private networks with VXLAN overlay, NAT gateways with bandwidth billing, subnet management and per-instance firewalls. Full IPv4 and IPv6.
Managed HAProxy with SSL termination, health checks, session stickiness and connection draining. Deploy in VPC or public mode.
PowerDNS, ClouDNS and Cloudflare wired into the panel, so zones follow instances automatically.
Every public instance, database, load balancer and control plane gets a real hostname automatically, with matching PTR records and a delegation wizard that verifies before publishing.
TLS certificates as account-owned objects, like AWS ACM: upload once or issue from Let’s Encrypt, reuse across listeners, serve multiple domains per listener via SNI, and get expiry alerts.
Load balancer rules match on host as well as path, so one load balancer serves several hostnames to different backends.
Managed clusters your customers can actually use, without a platform team.
Deep dive into Kubernetes →Control plane and worker nodes deployed as instances inside your VPC. kubectl and kubeconfig access out of the box.
Grow and shrink worker pools on demand, with a real autoscaler rather than a script.
Version upgrades roll through the cluster node by node, so workloads never go down.
Eight storage backends behind one wizard, from Ceph to ZFS.
Deep dive into Storage & Backups →Stream instance backups directly to S3, SFTP, FTP or any rclone remote. External snapshots, incremental chains, zero local staging.
Back up KVM and Proxmox VE instances to the same PBS destination: incremental, deduplicated, verified at rest, and kept after the instance is destroyed unless you choose to delete it.
Point-in-time rollback marks and DR exports for block volumes. Instant on-storage snapshots, in-place or new-volume restores.
S3-compatible buckets with self-service creation, access keys, per-bucket bandwidth tracking, quotas and hourly billing.
Ceph RBD, LVM thin and thick, ZFS thin and thick, QCOW2, NFS and Proxmox-native storage, onboarded through a guided wizard that probes the hypervisor before anything is saved.
Thin pools are watched every 30 seconds. When a pool nears exhaustion, only the at-risk instances are paused before disks corrupt, and they resume once the pool is fixed.
A hypervisor group carries a default backup policy every new instance inherits automatically, backups are throttled per node, and a dead node rebuilds from each instance’s latest backup chain onto a replacement.
Production database fleets your customers rent by the hour.
Deep dive into Managed Databases →Fully managed MySQL, MariaDB and PostgreSQL, deployed in seconds with hourly billing.
Read replicas, health monitoring and metrics dashboards for every deployment.
Point-in-time recovery and bring-your-own-S3 backups, restorable to the second.
Infrastructure as code, AI agents, and a developer experience your customers will pay for.
Deep dive into Automation & Developer Platform →65 resources and data sources, including MicroVMs, images and connectors as of v0.4.0, with full CRUD, import and dependency references, straight from the OpenTofu registry.
420 tools, including the MicroVM v3 family, with confirm-gated destructive actions, idempotent retries and async task convergence. Nothing to install.
Connect GitHub, GitLab, Bitbucket or Gitea. Builds detected automatically, PR preview environments, zero-downtime rollouts, automatic HTTPS.
Provision and manage the whole stack in plain language through guided chat. Operators get a parallel assistant for platform administration.
REST API, CLI, Terraform, MCP and webhooks. The user API, OpenTofu provider and MCP server address a location as location_id; the admin API keeps hypervisor_group_id.
A curated app catalogue customers can deploy to instances in one click.
Take payments directly with Stripe, Razorpay and PayPal. Customers top up credit, get tax-compliant PDF invoices by email, and request refunds.
Credit metering every 2 minutes across 12 resource types, from vCPU hours to NAT bandwidth.
WHMCS, Blesta, HostBill and Paymenter supported, so existing hosting stacks keep working.
The parts that keep a provider asleep at night.
Deep dive into Security & Operations →Role-based access for staff and customer teams, with white-label self-registration.
Node health checks every 30 seconds, alerting, and audit trails across the platform.
The control plane backs itself up and recovers automatically, with zero-downtime maintenance.
A real SSH shell in the browser, like Linode or DigitalOcean, plus noVNC and serial consoles. VPC-only guests are reached through the host, and credentials never touch the browser.
Customers sign in with Google, Microsoft or GitHub. Admin logins bind to your OIDC identity provider with strict subject binding and a logged break-glass path.
Prometheus scrape endpoints on the user API, so customers point their own Grafana at their instances, databases and load balancers with the token they already have.
The hypervisor agent detects ConfigServer Firewall and leaves host port management to it instead of contesting the INPUT chain, so security groups, VPC NAT and console proxies keep working.
A full multi-tenant helpdesk ships with the platform. No external tool, no separate login, and AI that reads the infrastructure before your staff do.
Deep dive into Support & Helpdesk →Customers open tickets against departments at four priorities. Pause-aware SLA timers track first response and resolution, with honest attainment reports.
The assistant reviews each new ticket and files a staff-only note: an assessment, a suggested reply your team loads with one click, and next actions. AI drafts, humans send.
A context card shows live status, metric sparklines, recent tasks and alerts for the exact resource the ticket is about. The same evidence grounds the AI triage.
Both sides reply on one thread. Staff-only internal notes are fail-closed: they can never leak into a customer-facing view.
Uploads are virus-scanned with ClamAV before acceptance, and images and PDFs preview inline. Stored on disk or your own S3 bucket.
Ticket events fan out to Slack, Discord, Telegram or webhooks, and threads update live over WebSocket. Support stays reachable even for suspended accounts.
30 days, every feature, no credit card, VirtConsole installs with the trial active by default. Just run the installer on your own hardware.
Debian 11/12/13 · Ubuntu 22.04/24.04/26.04 · Rocky / Alma / CentOS Stream 9/10 · run as root.