Frequently asked questions

Answers to common questions about Servor — the vault, agents, monitoring, status pages, the AI copilot, teams, billing and the API.

Quick answers to the questions people ask most about Servor. If you don't find what you need here, check the troubleshooting hub or the glossary.

What is Servor?

Servor is a server-management platform. You connect your servers, run commands from a browser terminal, watch them with monitors, publish a public status page, and let an AI copilot help you operate them. Everything sensitive is encrypted in your browser, so Servor can't read your secrets.

Do I need SSH to use Servor?

Only once. SSH is used for the initial setup — probing the machine, discovering it and installing the agent. After that, day-to-day work goes through the agent, not SSH. If the agent goes offline the server shows as unreachable; use Repair the agent to bring it back.

What is the vault, and why do I have to unlock it?

The vault is what keeps your sensitive data encrypted in your browser. On first sign-in you set a vault passphrase; you unlock the vault (enter that passphrase) once per session to run commands, use the web terminal, or use the AI copilot. Read-only viewers don't need to unlock. See Unlock your vault.

What happens if I forget my vault passphrase?

Servor cannot see or reset your passphrase — that's the point of a zero-knowledge vault. This is why you're given a recovery key (24 words) when you set up your vault. Save it somewhere safe. If you lose both the passphrase and the recovery key, encrypted data can't be recovered. See Account and security.

How do I add a server?

Go to Servers → Add a server. Servor gives you a one-line install command; copy it and run it in your server's shell. The agent installs, and the server shows as connected within seconds. There are 7 OS/arch targets and the agent auto-updates. Full walkthrough: Add a server.

Why does my server show as "unreachable"?

Because the agent isn't reporting in. All execution runs through the agent, so if it's stopped, crashed or the machine is offline, the server can't be reached. Run Repair the agent from the top bar or server settings, and see Connection issues for a full checklist.

Can Servor run dangerous commands?

A shared blacklist blocks known-dangerous commands (such as rm -rf /) across the terminal, command runner and AI copilot. The AI copilot also asks you to approve risky actions before they run. You stay in control.

What kinds of monitors can I create?

Nine types, including HTTP, TCP/port, ping, DNS, SSL certificate and keyword checks. Each monitor reports a status of up, degraded or down, tracks an uptime percentage, and runs on a check interval you choose. See Monitor types and Create a monitor.

How do I get notified when something goes down?

Set up alerts that trigger on status changes, then connect notification channels: Email, Slack, Discord or Webhook. Channels support a cooldown so you aren't flooded with repeated alerts. Configure both in Settings → Alerts.

Can I use a custom domain for my status page?

Yes. Point a CNAME at Servor's target and add the domain in your status page settings. The most common gotcha is Cloudflare: the orange-cloud proxy must be OFF (grey cloud / DNS-only), otherwise the certificate can't be issued. See Custom domain.

Can I make a status page private?

Yes. A status page can be public or password-protected. You can also let visitors subscribe for email/webhook updates. See Make it private and Subscribers.

What can the AI copilot do?

It's a chat copilot with a single tool: run_command. It works in three modes — Ask (read-only advice), Plan (proposes each action for your approval) and Auto (runs on its own, except destructive actions, which still need approval). See Ask, Plan, Auto and Get started with the copilot.

How is the AI copilot billed?

By credits, based on real token usage. When your balance runs low you buy top-up packages. See Credits and usage and Credits and top-ups.

What's the difference between the roles?

There are four: owner, admin, member and viewer. Viewers are read-only; members have write access and can use the terminal and AI but can't create servers, monitors or status pages; admins manage infrastructure, billing and the team; owners can do everything, including deleting the team. See Roles and permissions.

I invited a teammate but they can't see anything. Why?

New members can join instantly, but access to encrypted data requires a key grant. That happens automatically as soon as any teammate with an unlocked vault is online. If your teammate sees a "waiting for a teammate to come online" screen, have someone unlock their vault and access unlocks right away. See Invite teammates.

Troubleshooting — fix common problems
Glossary — plain-English definitions
Quickstart — from empty account to running command