Managing servers

Managing servers

When the JSON source is active, Settings → "Add / edit servers…" opens an editor to manage your connection list from a GUI — no hand-editing of JSON required.

What you can do

• Add a server with a form: Name, Group, User, Host or IP, Port, and an optional remote command.
• Edit any entry — including renaming it or moving it to another group.
• Reorder entries within a group by dragging them (use the ≡ handle / drag the row).
• Delete entries; a group with no entries left is removed automatically.

Entries with the same name (e.g. two "AdGuard" with different IPs) are kept distinct — editing, deleting or reordering one never affects the other.

Fields

Field	Notes
Name	Display name in the menu (required). Duplicate names are fine.
Group	Type a new group or pick an existing one. Empty → "Servers".
User	Optional; becomes user@host.
Host or IP	Required (in connection mode).
Port	Optional; omitted when 22.
Remote command	Optional. Runs on the server with a TTY: ssh -t user@host <command> — e.g. htop, tail -f /var/log/syslog. Empty = interactive shell.
Raw custom command	Toggle. Runs verbatim and ignores user/host/port — e.g. ssh -J jump root@10.0.0.5.

Running a command on a server

To pin a one-click action (monitor, restart, log tail), fill in Remote command:

• It keeps your User / Host / Port and runs ssh -t user@host '<command>'.
• The -t gives a TTY, so interactive tools like htop render correctly.
• Leave it empty for a normal interactive shell.

For something the connection fields can't express (jump hosts, tunnels), flip on Raw custom command and write the whole line yourself.

Safety

Host, IP and user are validated: values containing whitespace or shell metacharacters are rejected with an inline message. On top of that, every connection target is shell-quoted when the ssh command is built, so server data can't inject shell commands. See also Table import → Safety.

Every change writes the JSON file and snapshots the previous version — see Backups.