After the order: how an AI agent runs its own VPS over MCP

Description and use

An agent automates VPS lifecycle via MCP, handling status, metrics, power, plan changes, balance renewal, and support tickets, while destructive actions and balance funding need human input.

Problem

An AI agent must manage a VPS after ordering it, handling SSH readiness, monitoring metrics, renewing balance, and escalating issues.

Outcome

The agent can autonomously manage most VPS lifecycle tasks via MCP, but destructive actions and balance funding still require human intervention.

Workflow

• Order VPS via order_vps
• Poll get_vps_status until status is active and access.ssh_ready is true
• If no SSH key, call reveal_password after boot
• Monitor metrics with get_vps_metrics daily
• Use power_vps to start/stop/reboot as needed
• Rename with set_hostname
• Renew balance by checking get_balance and topup_balance when low
• Pay invoices with pay_invoice after top-up
• Handle plan changes with get_upgrade_options and change_plan
• Escalate issues with create_ticket and reply_ticket
• Delegate service to a human operator with delegate_service

Evidence

order_vps returns as soon as the order is paid from the balance. The server is created in the background, usually in about a minute.

Source

A well-behaved agent waits for access.ssh_ready instead of hammering port 22.

Source

If the order had no SSH key, the root password is waiting. reveal_password hands it over exactly once, and only after the VM has finished booting.

Source

get_vps_metrics returns CPU, memory, network and disk as time series for the last hour, day, week or month.

Source

power_vps does start, stop and reboot. One detail from testing: if the server isn't active, the tool returns a 409 with a hint to check the status first.

Source

When the balance is low, topup_balance produces a crypto checkout link: USDT, USDC or ETH on Base, Ethereum or Polygon.

Source

Destructive actions are deliberately awkward. reinstall_vps and an immediate cancel_service both require the server's exact hostname as confirm, and their descriptions tell the agent to check with a human before calling.

Source