Getting started

First 30 minutes

Do these steps in order. Each one unblocks the next, and by the end you have a node, delivery settings, a plan, a real account and a working bot.

The order that works#

1
Secure the panel. autoscript reset-pass to set your own password, then autoscript set-path if you want a different secret slug. Save both.
2
Set your brand and domains. Settings → General: brand name, panel domain, VPN hostname, timezone and currency display.
3
Join a node (or start with local-01). On the Nodes page create the node row, copy the generated install command, run it on the worker VPS, and wait for the provisioning progress to reach done. See Nodes.
4
Define delivery. Add a host entry (hostname + TLS/plain port sets) and, if you front traffic with Cloudflare, a CDN endpoint. See Hosts & CDN.
5
Create one test account per protocol. One SSH and one VLESS account is enough to prove the node, the ports and the certificates. Import the config into a client and connect.
6
Check enforcement. Connect twice from two devices on an account with a device limit of 1 and confirm the second session is cut. See Live connections.
7
Build a plan. Duration, data quota, device limit, protocols and price. See Plans.
8
Turn on payments. Add payment instructions / QR details and decide whether orders are auto-approved or manually reviewed. See Invoices & payments.
9
Connect the bot. Paste the Telegram token, set the trial policy and your admin chat id, then message the bot and complete one purchase yourself. See Telegram bot.
10
Set the SSH banner. Customise it in Settings → SSH banner with the live preview, then verify on a real SSH login. See SSH banner.
11
Turn on operations. Take a backup, confirm alert delivery, and note the release you are on: autoscript backup then autoscript release.

A 5-minute smoke test you can repeat#

panel VPS

autoscript status          # all services active, release known, install official
autoscript logs            # tail the stack for obvious errors
autoscript backup          # snapshot config + database

Nodes page: every node online, heartbeat age under a minute.
Connections page: your test session appears with the right source IP and protocol.
Accounts page: the test account shows non-zero usage after browsing through the tunnel.
Alerts page: no unacknowledged critical alerts.

Do this before every announcement

Run the smoke test right after any update or node change. It catches certificate, port and Xray-config problems before your users do.

Early pitfalls#

Panel host proxied through Cloudflare — TLS issuance and node RPC both fail. Keep it DNS-only.
Node joined but offline: the agent port is IP-locked to the panel’s addresses; a panel IP change requires re-running the node firewall step.
Client cannot connect on 443 while 8443 works: something else on the node already holds 443, or the CDN is stripping the transport.
Usage stays at zero: the account was created before the node finished provisioning — re-sync the node from the node detail page.
Bot silent: the token is set but the process is not running. Check systemctl status autoscript-bot.

Accounts

The daily-driver module.

Nodes

Add capacity properly.