Platform, agent, security, billing
Frequently asked questions
Answers about the platform, the agent, security and billing. If something here is vague, that is a bug. Tell us.
Platform
What Provibr is, and what it can drive today.
What is Provibr?
A server management platform for people who run their own infrastructure. You keep your hypervisors, your network and your data centre contracts; Provibr gives you and your team one panel to drive them from, and the machinery to hand capacity to customers.Which platforms can it manage today?
Hypervisors for virtual machines, control panels for hosting accounts and game panels for game servers. Dedicated servers, network gear and power distribution are on the roadmap. The integrations page carries the exact status of every provider, live or roadmap, so this answer can never quietly fall behind.Do I have to install anything on my hypervisor?
No. Provibr talks to the Proxmox API with a token you create, from an agent running on a separate Linux host in the same network. Nothing is installed on the hypervisor itself and no Proxmox port needs to be reachable from outside.Can I use my own IP addresses?
Yes, and that is the intended way to work. You register your prefixes, their gateways and the usable range in the platform's address management, and machines get an address from the pool you point their plan at.Which languages does the panel speak?
The full list is in the language switcher at the top of every page: 9 languages today. The panel, this site and the messages your agent generates are all translated, so the same event reads correctly for everyone on your team. The command-line tool and everything you paste into a shell stay in English on purpose.Is there an API?
Yes. There is a public REST API for servers, agents, integrations, addresses, plans, billing and webhooks; you create a key on your own profile page, and it carries a scope per part of the panel and never more than your own permissions. Because your agent connects outbound, an endpoint that changes a machine accepts the order and hands you something to follow rather than answering with the result. The documentation describes every endpoint and links to the OpenAPI 3.1 document.
Agent
The piece that runs inside your network.
How does the agent connect?
Outbound only. It opens a mutually authenticated TLS connection to the platform and keeps it open with a heartbeat. There is no inbound port, no port forward and no VPN, which is also why the agent works fine behind NAT.What does the agent run on?
A 64-bit x86 Linux host with systemd. It is a single static binary with no runtime dependencies, running as its own unprivileged service account with a hardened unit file.How do I install it?
You create the agent in the panel, which gives you a one-line command to run as root on the host. It checks the architecture, downloads the binary, verifies its checksum, creates the service account and directories, fetches the licence file, enrolls, and starts the service. Running it again on an existing host only refreshes the binary and the unit.How do updates work?
The panel shows you when an agent is behind and you press a button, either for one agent or for a selection at once. The new binary is downloaded over the existing connection, its signature and checksum are verified before anything is staged, and a root-owned script verifies them again before the binary is installed and the service restarted.What happens if an agent goes offline?
The panel marks it offline after the heartbeat stops, with a short grace period so a restart does not spam your activity log. Anything the agent could not deliver is replayed when it reconnects, in order, and a command is never executed twice. Machines it manages keep running. The agent is a control channel, not a dependency of your guests.Can I run more than one agent?
Yes, as many as you like, at no extra cost. One per location or per network segment is the usual pattern, and each carries only the credentials for the infrastructure it manages.
Security
Where your credentials live and how traffic is protected.
Where are my hypervisor credentials stored?
On the agent host, encrypted in a vault whose key is mixed with that machine's own identity. They travel through the platform once, in transit, to reach the agent. The audit record of that transfer keeps the field names and never the values, and the platform cannot reconstruct the token afterwards.How is the traffic protected?
TLS 1.3 with mutual authentication. Enrollment and the working session use separate endpoints, and on the session endpoint the platform pins the agent's certificate fingerprint, so a stolen certificate without the matching key gets nowhere.What happens if I revoke an agent?
The session is closed and the agent wipes its own identity and credential vault. From that moment it holds nothing useful: no certificate, no token, no way back in without a fresh licence and enrollment.Are the agent releases signed?
Yes, with a signature over the raw binary. It is verified by the agent before staging and again by a root-owned script before the binary is executed, against a public key that only root can replace. An unsigned or altered release is refused, even if its checksum matches.Can other customers see my data?
No. Every query, every realtime channel and every event is scoped to one organisation, and the realtime connection is authorised with a single-use ticket that is bound to your organisation before the socket opens.Is there an audit trail?
Yes. Commands, their results, agent connections and disconnections, integration status changes, address assignments, package changes and billing events all land in an activity log you can page through and filter. It is kept, not rotated away after a day.
Billing
What you pay for and what happens at the edges.
How is the price calculated?
Per resource in your package, per month, prepaid. A package with ten virtual machines costs ten times the per-machine rate whether three of them are running or all ten. The price calculator on the pricing page does exactly the same sum the platform does on an invoice.Is there a trial?
Yes, 14 days, with the whole panel and without payment details. When it ends nothing is charged; the platform just stops letting you start machines until a payment method is in place.How do I pay?
Through Stripe, with the methods Stripe offers in your country: iDEAL, Bancontact and card among them. The first payment sets up a mandate and every month after that is collected automatically.Do I get invoices?
Yes. Every collected payment produces a numbered invoice with your details, the lines it covers and the VAT. The panel shows the matching document and your full payment history, ready to print.What if I go over my limit?
You are told, in the panel and in your notifications, and you get a grace period to remove machines or move up a package. If nothing changes, the newest machines above the limit are stopped, never the oldest. Once you are back within the limit you start them again yourself.Can I change package?
Yes, and the change takes effect immediately. The difference is prorated over the rest of the period: an increase is added to your next invoice, and a decrease comes back as credit.
Next steps
Still unanswered?
The documentation goes into the parts that need commands and screenshots rather than a paragraph.
Start with one host and one agent
Create an account, install the agent with a single command and connect your first Proxmox host. The trial runs for 14 days and asks for no payment details.