Gryt CLI

The terminal manager for self-hosted Gryt servers.
Creates a server with working voice and uploads, then starts, stops, configures and updates every server on the machine.

Install

curl -fsSL https://get.gryt.chat | sh

The script picks the build for your platform, checks it against the release checksums, and installs to /usr/local/bin if that's writable or ~/.local/bin if not. GRYT_VERSION installs a specific tag instead of the newest release, and GRYT_INSTALL_DIR changes where the binary lands.

It doesn't cover Windows. Download the .zip from the releases page instead.

From source, with Go 1.25 or newer:

go install github.com/Gryt-chat/cli/cmd/gryt@latest

Docker Desktop or Docker Engine with the Compose plugin is required to start a generated deployment. Profile creation and .env generation work without it.

Full documentation: docs.gryt.chat/docs/cli.

Keyboard map

Key	Action
↑ / ↓ or k / j	Select server
n	New server wizard
e	Edit selected server
c	Change the settings the server keeps in its own database
enter	One server, with its addresses grouped by who they're for
s	Start
x	Stop
r	Restart
l	Recent logs
D	Remove the server, once you've typed its id
g	Refresh health
u	Update, when one is available
q	Quit

The wizard uses Enter to advance/save, Shift+Tab to move back, arrow keys to change a choice, and Esc to cancel.

Creating and starting a server from a script

gryt create --name my-server --port 5000 --yes
gryt start my-server

gryt create takes every flag the wizard asks for, and previews what it would create without writing anything until --yes is added:

Flag	Same as	Default
--name	Server name	required
--host	Bind address	0.0.0.0
--port	Port	the first port nothing else on the machine holds
--security	Security level (strict, balanced or community)	balanced
--voice-seats	Voice seats	0 (no limit)
--proxy-hops	Trusted proxy hops	0
--domain	The typed address in "Where will people connect from?"	none
--lan	Ticking this machine's own addresses in the same question	off

--domain takes a ws:// or wss:// address, and can be repeated or given as a comma-separated list. Without --domain or --lan, the server is only reachable at ws://localhost:5005 — the same as leaving every box but "This machine only" unticked in the wizard.

gryt start <server> brings up the shared voice server, then the named one. It's the same two steps behind the manager's s key, and it exits non-zero if Docker isn't available or the server was never created.

Both exit 0 on success and non-zero otherwise, so a script can chain them with gryt pull and gryt remove without driving a terminal.

Files

By default, profiles live below the platform user config directory:

gryt/
└── servers/
    └── my-server/
        ├── profile.json
        ├── .env
        ├── admin.env
        ├── compose.yaml
        └── data/

Set GRYT_CONFIG_DIR to use another root. Profile directories and files are created with private permissions where the operating system supports them.

admin.env holds the management token, apart from .env so that one stays safe to paste into a report. On Linux, data/ belongs to uid 1001, the user the server runs as. A data-init container hands it over each time the server starts.

Changing a running server

Some of a server's settings live in its database rather than its environment: who may join, whether it advertises itself over mDNS, whether the LAN is open, and what it does about profanity. c changes those on a running server, through the local management API the server publishes on loopback.

That needs server 1.5.0 or newer. Against anything older the screen says the server has no management API and points at gryt pull.

Everything else lives in the generated .env and takes effect on restart. The settings screen and gryt env both label which is which.

Moving a server onto a newer image

gryt pull my-server

That pulls the image the release channel points at, recreates the server and its image worker, and prints the version it was on next to the version it's on now. It stops before doing anything when the server isn't running, or when the server already runs the newest release. --force pulls without checking.

Restarting never pulled anything. docker compose restart reruns the container you already have, so the manager's version line names this command instead.

The voice server is a compose project of its own, shared by every server on the machine. gryt pull --shared moves it. Uploads live in each server's own data folder. Servers set up before that still keep theirs in a MinIO in the same shared project, and it stays for as long as one of them uses it.

Checking automatically instead of by hand

gryt pull --auto on

Installs the same nightly systemd timer Compose self-hosters use (ops/deploy/auto-update in the gryt repo), pointed at the shared voice server and every server this machine runs instead of gryt-<stack>-<service>. gryt pull --auto off removes it, and gryt pull --auto status says whether it's running. Re-run on after adding a server, because it writes the container list each time. Linux with systemd only, and it asks for sudo to install the unit files.

Removing a server

gryt remove my-server

Says what goes, asks you to type the server's id, then takes its containers down and deletes its folder, database and uploads included. --yes skips the question. The last server on the machine takes the voice server, the gryt network and shared/ with it. The images stay, and it prints the command that removes them.

Development

go test ./...
go vet ./...
go run ./cmd/gryt

Issues

Please report bugs and request features in the main Gryt repository.

Sponsors

What sponsoring pays for, the tiers, and everyone who has sponsored: gryt.chat/sponsors. To sponsor: GitHub Sponsors.

The list itself lives in the Gryt README, in one place rather than ten, so it cannot fall out of step across repositories.

License

AGPL-3.0 — Part of Gryt