← Back to Getting Started
OpenClaw getting started

First setup and onboarding

Use this page when OpenClaw is installed and you want to understand what onboarding does, what a normal first session should feel like, and how to tell whether setup basically worked.

The bridge Onboarding is what turns “the CLI exists” into “the system is configured enough to use.”
The next command

Run onboarding after install

openclaw onboard --install-daemon

Installing OpenClaw is only the first half. The next goal is to turn that install into a setup that actually feels usable.

Onboarding is the guided setup flow for:

  • model and provider auth
  • gateway configuration
  • first working local baseline
  • managed startup path when appropriate
Healthy first setup path

What normal looks like

  1. install OpenClaw
  2. run onboarding
  3. confirm the gateway is running
  4. open the dashboard
  5. send or attempt the first simple message or task

That sequence gives you believable proof that the system is not only installed, but actually usable.

Believable first run

What success should look like

Verify

Check the gateway and dashboard

openclaw gateway status
openclaw dashboard

What you want to see: the gateway reports as running, the dashboard opens, and the first interaction feels grounded rather than generic.

First questions

Ask one small local question

Good first checks: summarize the workspace, find the main config or docs file, or locate the most important startup rules.

If it feels wrong

Use evidence-first diagnosis

openclaw status
openclaw logs --follow
openclaw doctor

This is usually a better path than blind config edits if onboarding finished but the system still feels broken.

Where to go next

Keep moving through the newcomer path