A laptop or home server
- Sleeps or loses power
- Changing home IP address
- Exposes your home network
×OpenClaw on Nxeon
OpenClaw hosting on your own VPS.
OpenClaw is an open-source AI agent you run yourself. It needs a machine that stays on: a small always-on VPS with full root fits. From $8.88/mo for your first term, then renews at $12.78/mo.
- Full root
- Always on
- Snapshots
openclaw
# on your Nxeon server, after SSH
node --version # 24.16+ or 26.1+
# then follow the install steps in
# OpenClaw's own documentation
openclawyour agent gateway
startingsystemdkeeps it running 24/7
startingyour first setup
- Deploying Ubuntu 24.04 LTS
- SSH in as root
- Install Node.js 24.16 or newer
- Install OpenClaw from its docs
- Run it under systemd
Why a VPS for an AI agent?
An agent that answers you on chat apps has to be running when you message it. A laptop sleeps and a home connection drops. A small VPS stays on, has a fixed IP, and keeps your keys, memory and logs on a server only you can log into.
A Nxeon VPS
- Always on with a fixed IP
- Isolated from your home network
- Snapshots before every upgrade
- Resize as your use grows
OpenClaw isn't a one-click image on Nxeon, and we won't pretend otherwise. You deploy a clean Ubuntu 24.04 LTS server, SSH in and install it by following OpenClaw's own documentation. Nxeon provides the server: NVMe storage, a dedicated IPv4, snapshots, backups and DDoS protection. The model API you connect it to is a separate cost that you pay to that model provider.
What the server gives you
Full root
Install Node.js and the agent exactly the way its documentation describes, and change anything later.
Runs 24/7
Run it as a systemd service so it restarts after a reboot or a crash.
Snapshots before upgrades
Take an instant snapshot before you update, and roll back if the new version misbehaves.
Dedicated IP
A dedicated IPv4 is included, which is useful for allow-lists and webhooks.
DDoS protection and firewall
Keep the gateway private: allow only the ports you need through the managed firewall.
Room to grow
Resize CPU and RAM later without rebuilding if you add channels or browser automation.
Setup outline
- 1
Create your account
Sign up and open the deploy wizard.
- 2
Deploy Ubuntu 24.04 LTS on a 2 GB Starter or larger
Choose the region closest to you. Your server is online in about a minute.
- 3
Install Node.js
OpenClaw's install documentation asks for Node.js 24.16+ or 26.1+ (26 recommended) as of 2026-10-04. Its official installer is curl -fsSL https://openclaw.ai/install.sh | bash. Check the current requirement before you install.
- 4
Install and run OpenClaw
Follow the project's install instructions, then run it under systemd so it stays up. Keep API keys in environment files readable only by root.
Pick your size
Project documentation and community guides describe roughly 2 GB of RAM as a practical floor for a basic gateway that uses a hosted model, with more for several channels or browser automation. Start with Starter (2 GB) or Standard (4 GB), and resize later without a rebuild. Check OpenClaw's current docs, because requirements change between versions.
Starter
$8.88$12.78
First term, then $12.78 a month
- CPU1 vCPU
- Memory2 GB
- NVMe25 GB
- Bandwidth1 TB
Standard
$12.99$18.70
First term, then $18.70 a month
- CPU2 vCPU
- Memory4 GB
- NVMe50 GB
- Bandwidth2 TB
Pro
$20.99$30.22
First term, then $30.22 a month
- CPU3 vCPU
- Memory8 GB
- NVMe80 GB
- Bandwidth4 TB
Max
$32.99$47.50
First term, then $47.50 a month
- CPU4 vCPU
- Memory16 GB
- NVMe120 GB
- Bandwidth8 TB
OpenClaw hosting, answered
Anything else? Our support team answers around the clock.
Is there a one-click OpenClaw image?
No. You deploy a clean Ubuntu server and install OpenClaw yourself from its documentation. We would rather say that than imply a button that does not exist.
Does the VPS include the AI model?
No. Nxeon provides the server only. If OpenClaw is connected to a hosted model, that model's usage is billed by its own company. Running a model locally needs far more RAM than a small VPS has.
How much RAM do I need?
Guides describe about 2 GB as a floor for a basic setup using a hosted model, with more for multiple channels or browser automation. A 2 GB Starter is a reasonable start and you can resize any time.
Is it safe to expose?
Treat an agent with access to your accounts like any privileged service: keep it behind the firewall, expose only the ports you need, use HTTPS, and keep API keys out of shared files.
Can I move off Nxeon later?
Yes. It is your server with full root, so you can copy your configuration and data elsewhere whenever you like.
Your agent, on a server that's all yours.
Creating an account is free and takes no card details.