Skip to main content
This guide walks you through creating and deploying your first OpenClaw instance using ClawControl.

​
Prerequisites

Before you begin, you’ll need:
  • A cloud provider account (Hetzner Cloud or DigitalOcean)
  • An API key from your cloud provider
  • An AI provider account (Anthropic, OpenAI, OpenRouter, etc.)
  • A Telegram bot token (from @BotFather)
  • Your Telegram user ID for access control
Use the built-in templates to get started quickly. Templates pre-configure optimal settings for common use cases.

​
Creating Your Deployment

1

Launch ClawControl

Start ClawControl from your terminal:
clawcontrol
You’ll see the main menu with available commands.
2

Start the new deployment wizard

Type /new and press Enter to start creating a deployment.The wizard will guide you through several steps:Step 1: Choose a templateSelect a pre-configured template or start with a blank configuration:
> Blank configuration
  Hetzner + OpenRouter Kimi K2.5 [built-in]
  DigitalOcean + OpenRouter Kimi K2.5 [built-in]
Built-in templates are tested configurations that work well for most users. Choose “Hetzner + OpenRouter Kimi K2.5” for a cost-effective setup.
3

Name your deployment

Choose a unique name for your deployment:
Name: my-openclaw-server
  • Use lowercase letters, numbers, and hyphens only
  • Must be unique across your deployments
  • This name will be used for the VPS hostname
4

Configure cloud provider

If you selected a template, the provider is pre-selected. Otherwise, choose your provider:
> Hetzner Cloud - Recommended - US East
  DigitalOcean - NYC1 - New York
Enter your API key:The wizard validates your API key before proceeding.
ClawControl can securely save your API keys for future deployments. Press Y when prompted to save.
5

Select server size (DigitalOcean only)

If using DigitalOcean, choose your droplet size:
> s-1vcpu-2gb - 1 vCPU, 2GB RAM, 50GB SSD - $12/mo
  s-2vcpu-2gb - 2 vCPU, 2GB RAM, 60GB SSD - $18/mo
OpenClaw requires at least 2GB RAM to run reliably. Smaller sizes may cause memory errors.
6

Configure AI provider

Enter your AI provider details:Provider: Choose your AI service (Anthropic, OpenAI, OpenRouter, etc.)API Key: Enter your AI provider’s API keyModel: Specify the model identifier:
Anthropic:  claude-sonnet-4-20250514
OpenAI:     gpt-4o
OpenRouter: anthropic/claude-sonnet-4-20250514
Google:     gemini-2.0-flash
Groq:       llama-3.3-70b-versatile
7

Configure Telegram

Set up your Telegram bot:Bot Token: Get this from @BotFather
Format: 123456789:ABCdefGHIjklMNOpqrsTUVwxyz
Allowed User: Enter your Telegram user ID for access control
To find your Telegram user ID, message @userinfobot
8

Configure Tailscale (optional)

Choose whether to set up Tailscale for secure gateway access:
> Skip
  Configure Tailscale
Tailscale creates a private VPN so only you can access the OpenClaw gateway.
You can always add Tailscale later by editing your deployment configuration.
9

Confirm and create

Review your configuration and press Y to create the deployment:
Configuration Summary:

Name:         my-openclaw-server
Provider:     Hetzner Cloud
Location:     Ashburn, VA (US East)
Server:       CPX11 (2 vCPU, 2GB RAM)
AI Provider:  OpenRouter
Model:        moonshotai/kimi-k2.5
Channel:      Telegram

Press Y to confirm, N to go back
ClawControl creates the deployment configuration and saves it locally.

​
Deploying Your Instance

After creating the deployment configuration, you’re ready to deploy:
1

Start deployment

From the main menu, select your deployment and run:
/deploy
The deployment process begins automatically.
2

Monitor progress

ClawControl shows real-time progress as it:
[████████████████░░░░░░░░] 65%
Current: Installing Node.js

Deployment Log:
[14:23:45] Starting deployment: my-openclaw-server
[14:23:46] Creating VPS server...
[14:24:12] Waiting for server to start...
[14:24:45] Server created with IP: 159.89.123.45
[14:24:46] Connecting via SSH...
[14:25:01] SSH connection established
[14:25:02] Setting up swap memory...
[14:25:15] Updating system packages...
[14:27:34] Installing NVM...
[14:27:45] Installing Node.js...
The deployment includes 16 checkpoints:
  1. Server created
  2. SSH key uploaded
  3. SSH connected
  4. Swap configured
  5. System updated
  6. NVM installed
  7. Node.js installed
  8. pnpm installed
  9. Chrome installed
  10. OpenClaw installed
  11. OpenClaw configured
  12. Tailscale installed (if enabled)
  13. Tailscale authenticated (if enabled)
  14. Tailscale configured (if enabled)
  15. Daemon started
  16. Completed
3

Tailscale authentication (if enabled)

If you enabled Tailscale, you’ll be prompted to authenticate:
Confirmation Required

Tailscale needs authentication.

Would you like to open your browser to authenticate?
URL: https://login.tailscale.com/...

Press Y for Yes, N for No
Press Y to open the authentication page in your browser. After authenticating, the deployment continues automatically.
4

Optional interactive setup

Near the end of deployment, you may be prompted:
Your OpenClaw config has been applied and the daemon is running.

Would you like to also run 'openclaw onboard' interactively
for additional setup?
  • Press Y if you want to configure additional OpenClaw settings
  • Press N to skip (your agent is already configured and running)
If you choose Y, a terminal window opens with an SSH session.
5

Deployment complete

When deployment finishes successfully, you’ll see:
Deployment Successful!
Your OpenClaw instance is now running.

Connection Details
Server IP:      159.89.123.45
Tailscale IP:   100.64.1.23
Gateway Port:   18789

Next steps:
  /ssh  - Connect to your server
  /logs - View OpenClaw logs
  Gateway: http://100.64.1.23:18789/
Your OpenClaw instance is now live and ready to use!

​
Testing Your Deployment

1

Test Telegram bot

Open Telegram and start a conversation with your bot:
/start
Your bot should respond, confirming it’s running correctly.
2

Access the gateway

Open the gateway URL in your browser:
http://YOUR_TAILSCALE_IP:18789/
Or if Tailscale is not configured:
http://YOUR_SERVER_IP:18789/
If you’re not using Tailscale, your gateway is publicly accessible. Make sure to secure it or enable Tailscale.
3

Check deployment status

Back in ClawControl, run:
/status
Select your deployment to see:
  • Connection status
  • OpenClaw daemon health
  • Recent checkpoints

​
Troubleshooting

​
Deployment fails

If deployment fails, ClawControl automatically retries up to 3 times. If it still fails:
  1. Check the error message in the deployment log
  2. Run /status to see which checkpoint failed
  3. Run /deploy again - it resumes from the last successful checkpoint

​
Can’t connect to SSH

If SSH connection times out:
  1. Wait a few minutes for the server to finish booting
  2. Check your cloud provider’s firewall settings
  3. Verify the server was created in your cloud provider’s dashboard

​
OpenClaw not responding

If your bot doesn’t respond:
  1. Run /logs to check for errors
  2. Run /ssh to connect and check the service: 
    systemctl status openclaw
  3. Check your Telegram bot token and user ID are correct

​
Next Steps

Build docs developers (and LLMs) love

Get started for freeTalk to us