Skip to main content
The Rootly Setup Wizard is a guided command-line tool for getting a new Rootly workspace operational quickly. It walks you through creating a team, putting people on call, wiring up escalation and alerting, and creating a status page — then fires a real test alert (that pages whoever is on call) and opens a test incident, so you can see the whole flow end to end without clicking through every setup screen.
The Setup Wizard is for initial onboarding and configuration. For day-to-day incident and alert management from the terminal, see the CLI and Terminal UI.

​
Requirements

  • Node.js 18 or newer
  • A Rootly account (the wizard can hand you off to sign up if you don’t have one)
  • Sign-in via browser (OAuth) or a Rootly API token

​
Quick Start

Run the wizard with npx — no install required:
Sign in, then land on the main menu:
1

Sign in

Authorize with your browser (OAuth) or paste a Rootly API token. Your session is stored securely in your operating system’s keychain, so the next run takes you straight to the menu.
2

Choose a setup path

Pick Recommended setup to configure everything at once and see a live test alert and incident, or General setup to jump to any individual task.
3

Follow the guided steps

The wizard reuses anything that already exists, so it’s safe to re-run.

​
Signing in

From the sign-in screen you can:
  • Browser sign-in — authorize in your browser via OAuth.
  • API token — paste a Rootly API token. Recommended for the full experience (an admin, organization-wide key can complete every step).
  • Create a Rootly account — opens sign-up in your browser (account creation is web-only), then you return and sign in.
To create an API token:
  1. Log in to Rootly.
  2. Go to Organization Settings → API Keys.
  3. Click Generate New API Key, name it, and copy the token.
Your token or session is stored in your operating system’s keychain. You can also provide a token via the ROOTLY_TOKEN environment variable:
On exit, the wizard asks whether to keep the saved sign-in or delete it from your keychain (it keeps it by default). The fastest path. In one flow it will:
  1. Create a team (or reuse one you’re already on)
  2. Let you pick who joins the team and the on-call rotation
  3. Create an on-call schedule
  4. Create an escalation policy that pages the on-call schedule
  5. Add an alert source
  6. Create a public status page (and choose which components appear on it)
  7. Fire a test alert that pages the on-call person (a real call or text)
  8. Open a test incident (with a link to its Slack channel, if Slack is connected)
Before it runs, you can add and verify a phone number and connect Slack so the test alert actually reaches you. Anything that already exists is reused, so the flow is safe to re-run.

​
General setup

Jump to any individual task:

Teams & members

Create teams and add members from your directory.

On-call

Create on-call schedules and escalation policies.

Status page

Create a public status page and pick which components (services) appear on it.

Integrations

Connect Slack and alert sources (Datadog, Grafana, Sentry, PagerDuty, Opsgenie).

Verify

Send a test alert (pages on-call) or create a test incident.

Inspect

Review your current teams, schedules, and coverage.

MCP / IDE

Configure the Rootly MCP server for your editor or AI agent.

Chat with us

Talk to the Rootly team or book a demo.

​
MCP / IDE setup

The wizard can write Rootly MCP server configuration for your editor or AI agent — preview the config or apply it directly. It points the client at the hosted Rootly MCP server (https://mcp.rootly.com/mcp) and reads your token from the ROOTLY_TOKEN environment variable rather than embedding it inline.

​
Scripting (advanced)

Every setup step is also available non-interactively as a JSON-in / JSON-out action — handy for automation or AI agents. The interactive wizard requires a TTY; scripts and agents should use the action subcommand instead.
Each action returns exactly one JSON object on stdout:
Set ROOTLY_TOKEN to an organization admin API key for actions that read or write your workspace. Add "dryRun": true to any mutating action to validate and echo the call without executing it:
Actions that fire a test alert with paging enabled place a real call or text to whoever is on call, and creating a test incident opens a real incident in your workspace. Use a non-production workspace when experimenting.

​
Feedback & Support