OpenClaw is an open-source AI assistant framework that connects large language models — Claude, GPT-4o, Gemini, and others — to messaging platforms like Telegram, WhatsApp, Slack, and Discord. It runs as a lightweight Node.js process, which makes a small Linux VPS the ideal host. No GPU required, no desktop environment, no Mac Mini. Just a clean Ubuntu server and about 15 minutes of your time.
This guide walks you through every step: from ordering your VPS to verifying that your AI assistant is live and responding to messages. If you are still deciding between a VPS and a local machine, read our cost comparison of a VPS versus a Mac Mini for OpenClaw first.
Prerequisites
Before you begin, make sure you have the following ready:
- A Virtarix Cloud VPS — the $5.50-per-month plan (1 vCPU, 2 GB RAM, 30 GB NVMe) is more than enough. You can upgrade later if needed.
- An SSH client — Terminal on macOS and Linux, or PuTTY / Windows Terminal on Windows.
- At least one AI model API key — from OpenAI, Anthropic, Google AI, or another supported provider. OpenClaw itself is free (MIT license); you pay the model providers for usage.
- A messaging platform bot token — for example, a Telegram Bot Token from @BotFather, a Slack App token, a Discord Bot token, or WhatsApp Business API credentials. You need at least one to start.
Step 1 — Order and Access Your VPS
Choose a plan
Head to the Virtarix OpenClaw VPS page and select the Cloud VPS plan that fits your budget. The $5.50 plan handles a single-user assistant comfortably. If you plan to serve a team or run multiple bots, consider the 4 GB or 8 GB tier.
Connect via SSH
After provisioning, you will receive an IP address and root password by email. Open your terminal and connect:
ssh root@YOUR_SERVER_IP
Accept the fingerprint prompt and enter the root password. You are now logged in.
Step 2 — Secure Your Server
Before installing anything, harden the server. These steps take only a few minutes and protect you from the most common attacks.
Update system packages
apt update && apt upgrade -y
Change the root password
passwd
Choose a strong, unique password and store it in a password manager.
Create a non-root user
Running services as root is a security risk. Create a regular user with sudo privileges:
adduser openclaw
usermod -aG sudo openclaw
Set up SSH key authentication
On your local machine, generate a key pair if you do not already have one:
ssh-keygen -t ed25519 -C "openclaw-vps"
Copy the public key to your server:
ssh-copy-id openclaw@YOUR_SERVER_IP
Now log in as the new user to verify it works:
ssh openclaw@YOUR_SERVER_IP
Disable password authentication (optional but recommended)
Edit the SSH config:
sudo nano /etc/ssh/sshd_config
Find the line PasswordAuthentication yes and change it to PasswordAuthentication no. Save and restart SSH:
sudo systemctl restart sshd
Configure the firewall
sudo ufw allow OpenSSH
sudo ufw enable
Answer y when prompted. This allows SSH traffic and blocks everything else by default.
Install fail2ban
sudo apt install fail2ban -y
sudo systemctl enable fail2ban
sudo systemctl start fail2ban
Fail2ban monitors SSH login attempts and automatically bans IPs that fail repeatedly.
Step 3 — Install Node.js 22+
OpenClaw requires Node.js version 22 or later. The version in Ubuntu's default repositories is usually older, so use the official NodeSource script:
curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash -
sudo apt install nodejs -y
Verify the installation:
node --version
npm --version
You should see v22.x.x (or higher) and a recent npm version. If the node version is lower than 22, re-run the NodeSource setup script and install again.
Step 4 — Install OpenClaw
Install OpenClaw globally so it is available as a system-wide command:
sudo npm install -g openclaw@latest
Verify the installation:
openclaw --version
You should see the current OpenClaw version printed in your terminal.
Step 5 — Run the Onboarding Wizard
OpenClaw ships with an interactive onboarding command that configures your API keys, messaging integrations, and system daemon in one pass:
openclaw onboard --install-daemon
The wizard will prompt you for:
- AI model provider API keys — paste your OpenAI, Anthropic, or Google AI key. You can add multiple providers; OpenClaw will let you route messages to any of them.
-
Default model — choose which model handles messages by default (for example,
claude-sonnet-4-20250514orgpt-4o). - Messaging platforms — follow the prompts to connect Telegram, WhatsApp, Slack, Discord, or all of them. Each platform requires its own bot token or credentials.
-
Daemon installation — the
--install-daemonflag tells the wizard to create a systemd service so OpenClaw starts automatically on boot and restarts on crash.
Once the wizard finishes, OpenClaw is running. The systemd service is called openclaw.
Step 6 — Verify and Test
Check the service status
sudo systemctl status openclaw
You should see active (running) in green. If it shows failed, check the logs (see below).
Send a test message
Open the messaging platform you connected — Telegram, WhatsApp, Slack, or Discord — and send a message to your bot. You should receive a reply from the AI model within a few seconds.
View logs
If something is not working, check the OpenClaw logs:
journalctl -u openclaw -f
This streams the logs in real time. Press Ctrl+C to stop. Look for error messages related to API keys, network connectivity, or bot token validation.
Keeping OpenClaw Updated
OpenClaw is under active development. To update to the latest version:
sudo npm update -g openclaw@latest
sudo systemctl restart openclaw
It is good practice to take a VPS snapshot before upgrading, so you can roll back if anything breaks. Your Virtarix plan includes one snapshot slot for exactly this purpose.
Optional — Set Up a Reverse Proxy with HTTPS
If OpenClaw exposes a web interface or you want to serve webhook endpoints on a custom domain, set up Nginx with a free Let's Encrypt certificate.
Install Nginx and Certbot
sudo apt install nginx certbot python3-certbot-nginx -y
Allow HTTP and HTTPS through the firewall
sudo ufw allow 'Nginx Full'
Create an Nginx server block
sudo nano /etc/nginx/sites-available/openclaw
Paste the following configuration, replacing your-domain.com with your actual domain:
server {
listen 80;
server_name your-domain.com;
location / {
proxy_pass ;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_cache_bypass $http_upgrade;
}
}
Enable the site and get an SSL certificate:
sudo ln -s /etc/nginx/sites-available/openclaw /etc/nginx/sites-enabled/
sudo nginx -t
sudo systemctl reload nginx
sudo certbot --nginx -d your-domain.com
Certbot will automatically configure HTTPS and set up auto-renewal.
Troubleshooting
OpenClaw fails to start
Check the logs with journalctl -u openclaw -f. Common causes:
-
Node.js version too old — run
node --versionand ensure it is 22 or later. -
Missing API key — re-run
openclaw onboardand re-enter your keys. -
Port conflict — another service may be using the same port. Check with
sudo ss -tlnp.
Bot not responding to messages
- Verify your bot token is correct and the bot is not disabled on the messaging platform.
- For Telegram, make sure you sent
/startto the bot first. - For WhatsApp, confirm the webhook URL is reachable from Meta's servers (you may need the reverse proxy from the previous section).
- Check that your AI model API key is valid and has remaining credit.
High latency on replies
Response time depends mostly on the AI model provider, not your VPS. If replies are slow:
- Try a different or smaller model (for example,
gpt-4o-miniinstead ofgpt-4o). - Check your VPS resource usage with
htop— if CPU or RAM is maxed out, upgrade your plan. - Ensure your VPS is in a region close to the model provider's API endpoints.
Cannot connect via SSH
- Double-check the IP address and port.
- If you changed the SSH port or disabled password login, make sure you are using the correct port and key file.
- Check that UFW has not blocked your connection — if locked out, use the VPS console in your Virtarix client area.
Frequently Asked Questions
Is this a one-click install?
No. Virtarix provides a clean Ubuntu VPS with full root access. You install and configure OpenClaw yourself, which gives you complete control. The process takes about 15 minutes with this guide.
What VPS plan do I need?
The $5.50 plan (1 vCPU, 2 GB RAM, 30 GB NVMe) handles a personal AI assistant with moderate usage easily. For team use, heavy tool calling, or running additional services alongside OpenClaw, consider the 4 GB or 8 GB plan.
Can I run other services on the same VPS?
Yes. OpenClaw is lightweight. You can run Nginx, a personal wiki, monitoring tools, or other small Node.js apps on the same server without issues.
Does OpenClaw need a GPU?
No. OpenClaw sends prompts to cloud-based AI model APIs (OpenAI, Anthropic, Google). All inference runs on the provider's infrastructure. Your VPS only handles message routing and conversation state.
How do I migrate from a Mac Mini?
Copy your ~/.openclaw configuration directory to the VPS, install the same Node.js and OpenClaw versions, and re-run the onboarding wizard. Your bot tokens, API keys, and conversation history transfer over.
How much do the AI model APIs cost?
That depends on usage. Most casual users spend $5 to $20 per month on API calls. Set spending limits in your OpenAI, Anthropic, or Google dashboards to control costs. OpenClaw itself is free and open source.
What to Do Next
Your OpenClaw AI assistant is now running on a reliable, always-on VPS. Here are a few next steps to get even more out of your setup:
- Connect additional messaging platforms — the more channels you add, the more useful your assistant becomes.
- Experiment with different models — try Claude for long-form reasoning, GPT-4o for speed, or Gemini for multimodal tasks.
- Set up automatic backups — schedule VPS backups in your Virtarix control panel so you never lose your configuration.
- Explore tool use — configure OpenClaw to browse the web, call APIs, or search your files on command.
Need a VPS? Deploy your $5.50 OpenClaw VPS now and follow this guide to be up and running in 15 minutes.