manage-runners CLI
Manage runners, schedules and monitoring from your terminal, scripts and AI agents with the official manage-runners CLI.
manage-runners is the official command-line interface for Manage Runners. It talks to the same API as the dashboard, so runners you create in the terminal appear in the dashboard and the other way around. The CLI is open source under the MIT licence.
What you can do
| Task | Commands |
|---|---|
| Log in and manage profiles | config set-api, setup, auth login, auth status, auth logout, auth use |
| Choose an organization | org list, org use, --org |
| Find server types and prices | product list, product get, product locations |
| Manage runners | runner list, runner get, runner create, runner update, runner pause, runner resume, runner duplicate, runner delete, runner forget |
| Wait for runners | runner wait, runner watch, --wait |
| Read monitoring | runner metrics |
| Manage schedules | schedule get, schedule set, schedule enable, schedule disable, schedule delete |
| Check your account | profile show, subscription show, subscription plans |
| Diagnose problems | doctor, version |
| Update the CLI | version --check, self-update |
| Integrate with agents | commands, help --output json, setup skill, --agent |
Creating organizations, inviting members, managing billing and creating personal access tokens are only available in the dashboard.
Get started
- Install the CLI.
- Log in.
- Follow the task guides for runners and schedules, or browse the command reference.
Output formats
Every command supports three output formats:
--output human(the default): readable text.--output jsonor--json: one JSON document withok,command,data,metaandwarnings, orerrorwhen the command failed.--output jsonl: one JSON line per item for list commands, plus one line per warning.
Results and errors are written to standard output. Prompts go to standard error. The exit code tells you whether a command succeeded; see Exit codes.
Safety
The CLI treats changes that can cost money or delete data carefully:
- Commands that may create paid Hetzner resources (
runner create,runner resume,runner duplicate, andrunner updatewhen it recreates the server) ask for confirmation, or need--yes. - Destructive commands (
runner pause,runner delete,runner forget,schedule delete) ask for confirmation.runner delete,runner forgetandschedule deletealso need the exact runner ID. - The token the CLI creates when you log in can only read. Write, delete and forget permissions must be requested explicitly.
For agents and scripts
Use --agent for JSON output without prompts, discover commands with manage-runners commands --agent, and read the agent guide for the full contract.
