Best VPS for OpenClaw: How to Install & Run OpenClaw on Hostinger VPS (2026)
A complete guide to choosing the best VPS for OpenClaw and installing the OpenClaw AI agent framework on a Hostinger VPS. Covers server setup, full install commands, agent configuration, systemd auto-start, cost comparison across Hostinger, Hetzner, DigitalOcean, and Contabo, plus a VPS vs Raspberry Pi vs Mac Mini breakdown.
Why You Need a VPS for OpenClaw
OpenClaw is an open-source AI agent framework that lets you build and orchestrate autonomous agents from the command line. You can install OpenClaw on your laptop and experiment locally, but the moment you want your agents to run 24/7, respond to Telegram messages at any hour, execute scheduled heartbeat tasks, or stay online while your machine sleeps, you need a dedicated server. A VPS (Virtual Private Server) solves all of these problems.
A VPS for OpenClaw gives you a dedicated Linux environment with a static public IP address, guaranteed uptime, and full root access. Your OpenClaw AI agents stay running whether your laptop is open or not. Telegram bots respond instantly. Slack integrations never go offline. Scheduled heartbeat tasks fire on time, every time.
Running OpenClaw on a VPS also means you do not need to deal with port forwarding, dynamic DNS, or NAT traversal for webhooks. Your VPS has a public IP, so services like Telegram can reach your gateway directly. This is the standard production setup for anyone serious about deploying OpenClaw AI agents.
What to Look for in a VPS for OpenClaw
OpenClaw is lightweight. It does not run AI inference locally (unless you pair it with Ollama). The gateway process, agent routing, and channel management consume minimal CPU and memory. Here is what actually matters when choosing the best VPS for OpenClaw:
RAM: 2 GB minimum, 4 GB recommended
Node.js and the OpenClaw gateway need about 200-400 MB. Extra RAM gives headroom for multiple agents, session history, and any additional services you run alongside.
CPU: 1 vCPU minimum, 2 vCPU recommended
OpenClaw is not CPU-intensive since AI inference happens on provider servers. Two vCPUs help when multiple agents process requests concurrently.
Storage: 20 GB SSD minimum
The OpenClaw installation, Node.js, and agent configuration files use under 1 GB. Extra storage is for logs, session history, and any tools your agents interact with.
Network: Static IPv4 address
Required for Telegram webhook integration and any external service that needs to reach your gateway. Every reputable VPS provider includes this.
OS: Ubuntu 22.04 or 24.04 LTS
Ubuntu LTS is the most widely supported and documented choice. Debian, Fedora, and Alma Linux also work. Avoid Windows servers.
Best VPS Providers for OpenClaw (2026)
We tested OpenClaw on four popular VPS providers. Here is how they compare for running OpenClaw AI agents in production.
1. Hostinger VPS (Best Overall Value)
Hostinger VPS is the best value option for running OpenClaw in 2026. Their KVM-based virtual servers offer generous resources at the lowest price point among mainstream providers. The Hostinger VPS control panel (hPanel) makes server management straightforward even if you have never administered a Linux server before.
Key advantages of Hostinger VPS for OpenClaw: NVMe SSD storage across all plans, data centers in North America, Europe, Asia, and South America, a built-in firewall manager, weekly backups, and one-click OS reinstall. The KVM 2 plan at $5.99/month with 2 vCPU, 8 GB RAM, and 100 GB storage is the sweet spot for most OpenClaw deployments.
2. Hetzner (Best Performance per Dollar in Europe)
Hetzner offers exceptional raw performance from their data centers in Germany and Finland. Their CX22 plan (2 vCPU, 4 GB RAM, 40 GB SSD) costs around $4.85/month and handles OpenClaw with ease. Hetzner is particularly popular in the self-hosted community for its transparent pricing and reliable network. The downside is limited data center locations if you need low latency outside Europe.
3. DigitalOcean (Best Developer Experience)
DigitalOcean Droplets are the gold standard for developer-friendly VPS. Their $6/month Basic Droplet (1 vCPU, 1 GB RAM, 25 GB SSD) is the minimum viable option for OpenClaw, though the $12/month plan (2 vCPU, 2 GB RAM) is more comfortable. DigitalOcean excels in documentation, community tutorials, and a clean API. It costs more per resource than Hostinger or Hetzner, but the experience is polished.
4. Contabo (Most RAM for the Price)
Contabo is known for offering massive resource allocations at low prices. Their VPS S plan gives you 4 vCPU, 8 GB RAM, and 200 GB SSD for around $6.99/month. If you plan to run Ollama alongside OpenClaw for local inference, Contabo gives you the most RAM per dollar. The trade-off is slower disk I/O and a less modern control panel compared to Hostinger or DigitalOcean.
VPS Provider Comparison
| Provider | Plan | vCPU | RAM | Storage | Price/mo |
|---|---|---|---|---|---|
| Hostinger | KVM 2 | 2 | 8 GB | 100 GB NVMe | $5.99 |
| Hetzner | CX22 | 2 | 4 GB | 40 GB SSD | $4.85 |
| DigitalOcean | Basic 2GB | 2 | 2 GB | 50 GB SSD | $12.00 |
| Contabo | VPS S | 4 | 8 GB | 200 GB SSD | $6.99 |
Prices as of February 2026. Hostinger highlighted as recommended for the best balance of price, performance, and user experience.
Hostinger VPS Setup: Step by Step
This section walks through setting up a Hostinger VPS from scratch and preparing it for OpenClaw installation. If you already have a VPS from another provider, skip to the next section.
Step 1: Create a Hostinger VPS
Go to hostinger.com and select a VPS plan. The KVM 2 plan ($5.99/month) is the recommended starting point. During setup, choose Ubuntu 22.04 or 24.04 as your operating system. Set a strong root password or upload your SSH public key. Select the data center closest to your location.
Step 2: Connect to Your VPS via SSH
Once your Hostinger VPS is provisioned (usually within 60 seconds), find the IP address in your hPanel dashboard. Open your terminal and connect:
# Connect to your Hostinger VPS
ssh root@YOUR_VPS_IP_ADDRESS
# If you set up SSH key during creation:
ssh -i ~/.ssh/id_rsa root@YOUR_VPS_IP_ADDRESSStep 3: Secure the Server
Before installing anything, lock down your VPS with basic security measures. Create a non-root user, disable root password login, and configure the firewall.
# Update system packages
apt update && apt upgrade -y
# Create a non-root user
adduser openclaw
usermod -aG sudo openclaw
# Set up SSH key for the new user
mkdir -p /home/openclaw/.ssh
cp ~/.ssh/authorized_keys /home/openclaw/.ssh/
chown -R openclaw:openclaw /home/openclaw/.ssh
# Configure firewall (allow SSH and OpenClaw gateway)
ufw allow OpenSSH
ufw allow 18789/tcp
ufw enable
# Switch to the new user
su - openclawStep 4: Install Node.js
OpenClaw requires Node.js 18 or higher. The recommended way to install it on a VPS is through NodeSource or nvm.
# Option A: Install Node.js 20 via NodeSource (recommended for VPS)
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt install -y nodejs
# Verify installation
node --version # Should output v20.x.x
npm --version # Should output 10.x.x
# Option B: Install via nvm (more flexible)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.7/install.sh | bash
source ~/.bashrc
nvm install 20
nvm use 20How to Install OpenClaw on Your VPS
With Node.js installed on your VPS, you can now install OpenClaw globally. This section covers the full installation process, from the npm install command through initial configuration and first agent launch.
Install OpenClaw via npm
# Install OpenClaw globally
npm install -g openclaw
# Verify installation
openclaw --version
# Run the interactive configuration wizard
openclaw configureThe openclaw configure wizard will walk you through setting up your model provider, API key, default model, and communication channels. For a VPS setup, pay attention to the gateway host setting.
Configure OpenClaw for VPS
After the initial wizard, apply these VPS-specific settings. The key difference from a local install is binding the gateway to 0.0.0.0 so it accepts connections from external services like Telegram webhooks.
# Set your AI model provider API key
openclaw models auth paste-token --provider anthropic
# Set the primary model
openclaw config set agents.defaults.model.primary claude-sonnet-4-20250514
# Set a fallback model (optional but recommended)
openclaw config set agents.defaults.model.secondary gpt-4o-mini
# Bind gateway to all interfaces (required for VPS)
openclaw config set gateway.host 0.0.0.0
openclaw config set gateway.port 18789
# Set up Telegram channel (optional)
openclaw config set channels.telegram.bottoken YOUR_TELEGRAM_BOT_TOKEN
openclaw config set agents.defaults.channel telegram
# Enable heartbeat for scheduled tasks
openclaw config set agents.defaults.heartbeat true
openclaw config set agents.defaults.heartbeat.interval 3600Security note: When binding to 0.0.0.0 on a public VPS, make sure your firewall only allows traffic on port 18789 from trusted sources, or place the gateway behind a reverse proxy like Nginx with HTTPS.
Start the OpenClaw Gateway
# Start the gateway
openclaw gateway start
# Verify it is running
curl http://localhost:18789/health
# Check gateway logs
openclaw gateway logsKeep OpenClaw Running with systemd
When you close your SSH session, any process started in that shell will terminate. To keep OpenClaw running permanently on your VPS, create a systemd service. This also ensures the gateway restarts automatically after a server reboot or crash.
# Create the systemd service file
sudo nano /etc/systemd/system/openclaw.service
# Paste this configuration:
[Unit]
Description=OpenClaw AI Agent Gateway
After=network.target
[Service]
Type=simple
User=openclaw
WorkingDirectory=/home/openclaw
ExecStart=/usr/bin/openclaw gateway start
Restart=on-failure
RestartSec=10
Environment=NODE_ENV=production
[Install]
WantedBy=multi-user.target# Enable and start the service
sudo systemctl daemon-reload
sudo systemctl enable openclaw
sudo systemctl start openclaw
# Check status
sudo systemctl status openclaw
# View logs
sudo journalctl -u openclaw -f
# Restart after config changes
sudo systemctl restart openclawWith systemd enabled, your OpenClaw gateway will survive SSH disconnects, server reboots, and process crashes. This is the standard way to run any long-running service on a Linux VPS.
Configuring Your First Agent on the VPS
Now that OpenClaw is installed and running on your VPS, it is time to create and register your first agent. Every OpenClaw agent is defined by a SOUL.md file that specifies its identity, rules, skills, and behavior.
Create an Agent Workspace
# Create a project directory for your agents
mkdir -p ~/my-agents/agents/content-writer
# Create the SOUL.md file
nano ~/my-agents/agents/content-writer/SOUL.mdWrite a SOUL.md
Here is a minimal SOUL.md to get your first agent running. It defines an SEO content writer agent.
# Content Writer Agent
## Identity
You are a senior SEO content writer. You produce clear,
well-structured blog posts optimized for search engines.
## Rules
- Write in English only
- Use short paragraphs (2-3 sentences)
- Include relevant keywords naturally
- Always include a meta description suggestion
## Skills
- Blog post writing
- SEO optimization
- Content editing
- Headline generation
## Channel
telegramRegister and Test the Agent
# Register the agent with OpenClaw
cd ~/my-agents
openclaw agents add content-writer \
--workspace ./agents/content-writer \
--non-interactive
# Test the agent with a message
openclaw agent --agent content-writer \
--message "Write a 200-word intro about AI agents"
# Restart gateway to pick up the new agent
openclaw gateway restart
# Verify the agent is registered
openclaw agents listOnce registered, your agent is accessible through whatever channels you configured. If you set up Telegram, you can now message your bot and it will route to your content-writer agent.
Optional: Nginx Reverse Proxy with HTTPS
For production OpenClaw deployments on a VPS, it is best practice to put the gateway behind an Nginx reverse proxy with SSL. This gives you HTTPS encryption, better security, and the ability to serve the gateway on port 443.
# Install Nginx and Certbot
sudo apt install -y nginx certbot python3-certbot-nginx
# Create Nginx config for OpenClaw
sudo nano /etc/nginx/sites-available/openclaw
# Paste this configuration:
server {
listen 80;
server_name your-domain.com;
location / {
proxy_pass http://127.0.0.1:18789;
proxy_http_version 1.1;
proxy_set_header Upgrade $http_upgrade;
proxy_set_header Connection 'upgrade';
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_cache_bypass $http_upgrade;
}
}
# Enable the site
sudo ln -s /etc/nginx/sites-available/openclaw /etc/nginx/sites-enabled/
sudo nginx -t
sudo systemctl reload nginx
# Get a free SSL certificate from Let's Encrypt
sudo certbot --nginx -d your-domain.comAfter this setup, your OpenClaw gateway is accessible via HTTPS on your domain. When using a reverse proxy, change the gateway host back to localhost: openclaw config set gateway.host 127.0.0.1
VPS vs Raspberry Pi vs Mac Mini: Which Is Best for OpenClaw?
There are three popular ways to self-host OpenClaw. Each has trade-offs depending on your budget, technical requirements, and whether you need a public IP or prefer local-only operation.
| Feature | VPS (Hostinger) | Raspberry Pi 5 | Mac Mini |
|---|---|---|---|
| Upfront Cost | $0 | $80-120 | $599+ |
| Monthly Cost | $5-10/mo | $3-5/mo (electricity) | $5-8/mo (electricity) |
| Public IP | Yes (included) | No (needs port forwarding) | No (needs port forwarding) |
| Uptime | 99.9% SLA | Depends on home power/network | Depends on home power/network |
| Telegram Webhooks | Works out of the box | Requires tunneling (ngrok) | Requires tunneling (ngrok) |
| Local Ollama | Possible (CPU only) | Yes (ARM, slower) | Yes (fast, Apple Silicon) |
| Setup Difficulty | Medium | Medium-High | Low |
| Best For | Production, 24/7 agents | Experimentation, offline | Power users, local AI |
For most users who want to run OpenClaw AI agents in production with Telegram or Slack integration, a VPS is the clear winner. The Hostinger VPS KVM 2 plan at $5.99/month gives you a hassle-free production environment. The Raspberry Pi is a great choice for tinkering and local-only setups. The Mac Mini is overkill for OpenClaw alone but makes sense if you also run local models.
Hostinger VPS Cost Breakdown for OpenClaw
Here is a detailed breakdown of Hostinger VPS plans and which one suits different OpenClaw deployment sizes.
| Plan | vCPU | RAM | Storage | Price/mo | OpenClaw Use Case |
|---|---|---|---|---|---|
| KVM 1 | 1 | 4 GB | 50 GB NVMe | $4.99 | 1-2 agents, basic setup |
| KVM 2 | 2 | 8 GB | 100 GB NVMe | $5.99 | Recommended: 5-10 agents |
| KVM 4 | 4 | 16 GB | 200 GB NVMe | $9.99 | 10+ agents, Ollama, heavy use |
| KVM 8 | 8 | 32 GB | 400 GB NVMe | $15.99 | Large teams, multi-project |
The total cost of running OpenClaw on a Hostinger VPS is the VPS plan plus your AI model API costs. For context, using Claude Sonnet as your primary model costs roughly $3 per 1 million input tokens. A typical agent handling 50 queries per day costs $1-3/month in API fees. So your total monthly cost for a production OpenClaw setup on Hostinger is approximately $7-10.
Troubleshooting OpenClaw on VPS
"Cannot connect to gateway from outside"
Make sure the gateway is bound to 0.0.0.0 and not localhost. Run openclaw config set gateway.host 0.0.0.0 and restart. Also verify your firewall allows port 18789: sudo ufw allow 18789/tcp.
"Gateway stops when I close SSH"
You need to run the gateway as a systemd service, not directly in your terminal. Follow the systemd setup section above. As a quick workaround, use tmux new -s openclaw to start a persistent terminal session, then detach with Ctrl+B, D.
"npm install -g openclaw fails with permission error"
If you installed Node.js via apt, global npm installs require sudo: sudo npm install -g openclaw. Alternatively, install Node.js via nvm, which installs to your home directory and does not require sudo for global packages.
"Telegram bot not responding"
Verify the bot token is set correctly: cat ~/.openclaw/config.json | grep bottoken. Make sure the gateway is running and reachable from the internet. Check logs with sudo journalctl -u openclaw -f to see if Telegram webhook connections are arriving.
"Out of memory on small VPS"
If your VPS has only 1 GB RAM, add a swap file: sudo fallocate -l 2G /swapfile && sudo chmod 600 /swapfile && sudo mkswap /swapfile && sudo swapon /swapfile. Add it to /etc/fstab to persist across reboots. For sustained use, upgrading to a 2+ GB plan is recommended.
Frequently Asked Questions
What is the best VPS for running OpenClaw AI agents?
Hostinger VPS is the best value option for running OpenClaw, starting at $5.99/month for 2 vCPU, 8 GB RAM, and 100 GB NVMe storage. It provides more than enough resources to run multiple OpenClaw agents 24/7. Hetzner and DigitalOcean are also strong alternatives, with Hetzner offering the best raw performance per dollar in Europe and DigitalOcean providing the simplest managed experience.
Can I run OpenClaw on a VPS?
Yes. OpenClaw runs on any Linux VPS with Node.js 18 or higher. A VPS gives your agents a permanent, always-on server with a static IP address. This means your Telegram bot, Slack integration, and scheduled heartbeat tasks all keep running even when your local machine is off. Any VPS with at least 1 GB RAM and 1 vCPU is sufficient for a basic OpenClaw setup with one or two agents.
How do I install OpenClaw on a VPS?
SSH into your VPS, install Node.js 18+ using nvm or your package manager, then run 'npm install -g openclaw' to install OpenClaw globally. After installation, run 'openclaw configure' to set up your model provider, API key, and channels. Finally, start the gateway with 'openclaw gateway start' and optionally set up a systemd service to keep it running after reboot. The full step-by-step commands are covered in this guide.
How much does a Hostinger VPS cost for OpenClaw?
Hostinger VPS plans for OpenClaw start at $5.99/month (KVM 2) with 2 vCPU, 8 GB RAM, and 100 GB NVMe SSD. This plan can comfortably run 5 or more OpenClaw agents simultaneously. The KVM 1 plan at $4.99/month with 1 vCPU and 4 GB RAM works for 1-2 agents. Hostinger frequently runs promotions that bring the price even lower for the first term. All plans include a dedicated IPv4 address and full root access.
Should I use a VPS or Raspberry Pi for OpenClaw?
A VPS is better for production use because it offers a static public IP, professional uptime guarantees, automatic backups, and no dependency on your home network or power. A Raspberry Pi is better for local experimentation, offline setups with Ollama, or when you want to avoid monthly costs. For agents that need to be reachable via Telegram or Slack webhooks from anywhere, a VPS is the recommended choice.
Do I need a VPS with GPU for OpenClaw?
No. OpenClaw itself does not require a GPU because it sends inference requests to cloud AI providers like Anthropic or OpenAI. A standard CPU-only VPS is all you need. The only exception is if you plan to run local models through Ollama on the same VPS, in which case a GPU instance from providers like Lambda or Vast.ai would be beneficial. For most users, a $5-10/month CPU VPS from Hostinger, Hetzner, or DigitalOcean is more than sufficient.
How do I keep OpenClaw running after I close the SSH session?
Use systemd to create a service that starts OpenClaw automatically on boot and restarts it if it crashes. Create a service file at /etc/systemd/system/openclaw.service, enable it with 'systemctl enable openclaw', and start it with 'systemctl start openclaw'. Alternatively, you can use pm2 (a Node.js process manager) or run the gateway inside a tmux or screen session. Systemd is the most reliable option for production VPS deployments.
Can I run multiple OpenClaw agents on one VPS?
Yes. A single OpenClaw gateway manages all your agents, so you only need one VPS regardless of how many agents you run. Each agent consumes minimal resources since the heavy computation happens on the AI provider side. A VPS with 2 GB RAM can handle 10+ agents comfortably. The main bottleneck is API rate limits from your model provider, not VPS resources.
Skip the Manual Config Writing
Generate your SOUL.md and all agent configuration files at crewclaw.com in seconds. Pick a role, connect tools, download your deploy-ready package, and scp it to your VPS.
Build Your Agent ConfigFree to generate. $9 one-time to download the full package.