How to Deploy OpenClaw (MoltBot/ClawdBot) on Linux VPS – Complete Guide
- January 26, 2026
- 2:42 am
- Tutorials
OpenClaw (formerly known as MoltBot or ClawdBot) is the evolution of the open-source personal AI assistant. While the name has changed to reflect its growing ecosystem, the mission remains the same: giving you an AI that runs entirely on your infrastructure.
Unlike standard chatbots like ChatGPT or Gemini which live in a browser tab, OpenClaw connects directly to your daily life via WhatsApp, Telegram, Discord, Signal, and Slack. It provides seamless, proactive interactions while ensuring your data stays under your control, not used to train public models.
Why Self-Hosting is the Future of AI
The “Mac Mini” Trap is a common mistake. We see many enthusiasts spending $600+ on hardware like a Mac Mini or Raspberry Pi to run their “Self-hosted AI”.
Why is this a mistake? Home servers suffer from residential internet latency, power outages, and noise. If your home connection drops for a second, your AI assistant goes offline. Furthermore, keeping a computer running 24/7 at home impacts your electricity bill significantly.
The smartest solution for a Node.js AI Agent like OpenClaw is a high-performance Linux VPS. It runs 24/7 in a secure datacenter with a 10Gbps connection, ensuring your bot replies in milliseconds.
Step 1: Choosing the Right VPS Architecture
OpenClaw is built on a modern architecture that requires Node.js v22 and sufficient RAM to handle the “Gateway” — the service that maintains the WebSocket connections with WhatsApp or Discord.
We recommend the High Performance range from RDP Monster, powered by AMD Ryzen™ 9 7950X. Why Ryzen? AI logic and Node.js are single-thread heavy. The high clock speed of the Ryzen 9 ensures that your bot processes context and generates answers faster than standard cloud servers.
Step 2: System Preparation & Dependencies
Before installing the bot, we need to prepare the Linux environment. OpenClaw handles multimedia (voice notes, images), so we need specific libraries.
Connect to your VPS via SSH:
Crucial: Install Node.js 22 and FFmpeg. FFmpeg is mandatory if you want your bot to listen to WhatsApp voice notes and transcribe them.
# Install Node.js 22 repo
curl -fsSL https://deb.nodesource.com/setup_22.x | bash –
# Install dependencies
apt install -y nodejs ffmpeg git
Step 3: Installation & Onboarding Wizard
OpenClaw simplifies the complex configuration of AI Agents through a CLI (Command Line Interface). We will install it globally.
openclaw onboard
* Gateway Mode: Select
Local. This tells OpenClaw to run the server on this VPS.* Auth: Use
Anthropic API Key (Claude 3.5 Sonnet is currently the best model for coding and reasoning) or OpenAI.* Install Daemon: Select Yes. This automatically sets up a systemd service to keep your bot running even if you close the terminal.
Step 4: Starting the Gateway (The Missing Link)
This is where most tutorials fail. The “Onboard” command creates the config, but it doesn’t start the engine. You need to launch the Gateway and ensure the firewall allows access to the Web Dashboard.
Open port 18789. On a secure Linux VPS, ports are closed by default. We use UFW (Uncomplicated Firewall) to allow traffic.
Start the Gateway in background mode:
Note: If you already selected “Install Daemon” in step 3, simply check if it’s running with openclaw health.
http://YOUR_VPS_IP:18789 in your browser. You should see the OpenClaw Control Panel, a visual interface to manage your bot’s memory and connections.
Step 5: Connecting Channels (WhatsApp & Discord)
OpenClaw is multi-channel. You can have the same AI assistant available on WhatsApp for personal texts and Discord for server management.
Option A: WhatsApp Integration
OpenClaw uses a library that simulates a Web Browser connection. Run this command to generate a QR code in your terminal:
Scan it immediately with your phone (Linked Devices > Link a Device).
Option B: Discord Bot
For Discord, you need to create an application in the Discord Developer Portal.
Important: You must enable “Privileged Gateway Intents” (Message Content Intent & Server Members Intent). Without this, your OpenClaw instance cannot read user messages.
Step 6: Security & Pairing
Since your VPS is on the public internet, you don’t want strangers using your API credits. OpenClaw implements a Pairing System (Allowlist). By default, it ignores everyone.
# 2. Check pending requests on the VPS
openclaw pairing list
# 3. Approve your specific number
openclaw pairing approve whatsapp YOUR_PHONE_NUMBER
Step 7: Features & Advanced Usage
Your AI is now alive on your high-performance VPS. Unlike basic chatbots, OpenClaw has Persistent Memory. It remembers context across days and sessions.
Essential Commands:
/think high: Activates “Reasoning Mode” (Chain of Thought). Perfect for complex coding tasks or math./new: Forces a context reset if you want to change topics completely./summary: Asks the bot to summarize the last 50 messages (great for group chats).
OpenClaw isn’t just a chatter. It supports an extensive Skills System. You can install plugins to let it:
– Search the live web (Perplexity/Google style)
– Read and summarize your emails
– Manage your Google Calendar
– Scrape websites for data
Ready to Deploy?
Don’t rely on unstable local hardware. Get a professional environment for your personal AI.
Deploy My AI AssistantPromo code: CLAWD10 (-10%)
Powerful Windows VPS Hosting
Enjoy seamless control and high-speed performance with our Windows RDP solutions. Ideal for managing your servers, running applications securely, and boosting your productivity from anywhere.
High-Performance Dedicated Servers
Need maximum control and power? Our Dedicated Servers offer unmatched performance for demanding tasks.
Frequently Asked Questions
How much does it really cost?
Why not host it on my own PC?
🔒 Is my data private?
Messages travel only between your VPS (which you control) and Anthropic's secure servers (to generate the response). No one else reads your conversations.
