How OpenClaw Onboarding Works
Learn about the initial setup flow and what gets configured.
What Is Onboarding?
Onboarding is the first-time setup process that configures OpenClaw for your use case. It walks you through connecting models, channels, and basic security settings.
Quickstart vs Manual Onboarding
Quickstart
Guided wizard with recommended defaults. Best for beginners.
openclaw onboard
Manual
Full control over every setting. Best for advanced users.
openclaw configure
What the Onboard Command Does
Running openclaw onboard walks you through:
- 1.Model setup — Connect OpenAI, Anthropic, or Ollama
- 2.Channel login — Connect Telegram, Discord, or WhatsApp
- 3.Security defaults — Applies hardened baseline settings
- 4.Daemon install — Sets up auto-start on boot
Install Daemon
The --install-daemon flag sets up OpenClaw to start automatically:
openclaw onboard --install-daemon
- • macOS — launchd service (runs at login)
- • Linux — systemd user service
- • Windows — Scheduled task
Pairing Flow
Pairing connects your devices (phone, laptop) to the gateway:
- 1.Run
openclaw pairon your computer - 2.Scan the QR code with your phone
- 3.Your device is now paired and can access the gateway
Gateway Authentication
OpenClaw supports three authentication modes:
Token Auth
Set a token in config. Clients pass the token when connecting.
Password Auth
Simple password-based authentication.
Device Auth
Pair devices via QR code. No persistent tokens needed.
Common Onboarding Mistakes
- ✗Skipping daemon install — Gateway doesn't start on boot
- ✗Not setting allowlists — Anyone can message your bot
- ✗Forgetting API key — Gateway can't connect to models
- ✗Wrong port — Clients connect to wrong port
Ready to start?
Run onboarding and get set up in minutes.
openclaw onboard --install-daemon