Should I install OpenClaw on my computer or a VPS?

For personal use, a local install is fine. For business automation that needs 24/7 uptime, a VPS is strongly recommended. When your laptop sleeps or loses power, OpenClaw stops — meaning missed follow-ups, dead automations, and lost leads. A VPS runs continuously in a datacenter with guaranteed uptime.

What are the system requirements for OpenClaw?

Minimum: 1 CPU core, 2 GB RAM, 10 GB disk, Node.js 22+, and an internet connection. Recommended for business use: 2+ cores, 4 GB RAM on Ubuntu 22.04 or 24.04. A Raspberry Pi 4/5 with 4-8 GB RAM also works well as a dedicated device.

How much does it cost to run OpenClaw on a VPS?

VPS hosting costs $5-24/month depending on the provider. Oracle Cloud offers a free tier with 4 ARM CPUs and 24 GB RAM. DigitalOcean and Hostinger offer one-click OpenClaw deploys starting at $5-6/month. Your only other cost is the AI API (Anthropic, OpenRouter, or use free local models via Ollama).

Is OpenClaw safe to install? What are the security risks?

OpenClaw is MIT-licensed, free, and open-source — fully auditable. The main risks are: giving an AI agent shell access and API tokens, and leaving the gateway port exposed on the internet. Best practices: run on a dedicated device or VPS (never your daily driver), bind to localhost, use Docker isolation, enable command allowlisting, and use Tailscale for remote access instead of exposing ports.

Can I install OpenClaw with Docker?

Yes, Docker is the recommended method for VPS deployment. Clone the repo, run the docker-setup.sh script, and Docker Compose handles everything. Volumes mount ~/.openclaw for config and ~/openclaw/workspace as the agent's sandbox. Docker provides container isolation which limits the blast radius if something goes wrong.

How do I connect OpenClaw to WhatsApp, Telegram, or Slack?

After installation, the setup wizard walks you through pairing. For Telegram: create a bot via @BotFather, paste the token, and pair via a code sent to your Telegram. For WhatsApp: scan a QR code or use the Business API. For Slack/Discord: add the OpenClaw app via OAuth. Our free setup guide includes step-by-step guidance for each platform.

How to Install OpenClaw on Mac, Windows, Linux, Docker, or VPS

To install OpenClaw (formerly ClawdBot): Run curl -sSL https://get.openclaw.ai/install.sh | bash on Mac or Linux. For Windows, use WSL2 first. For business use, deploy on a VPS with Docker for 24/7 uptime and security isolation. The installer handles all dependencies automatically — setup takes under 15 minutes.

This guide covers every installation method, compares local device vs VPS deployment, and includes security hardening, troubleshooting, and post-install configuration.

VPS Setup (Recommended)Local Device Install Device vs VPS Comparison Security Checklist Troubleshooting FAQ

Why Does Your Installation Method Matter?

OpenClaw is a proactive AI agent — it sends reminders, follows up on leads, monitors your systems, and executes tasks around the clock. If it goes offline because your laptop sleeps, your automations stop. For business use, where you run OpenClaw determines your uptime, security, and reliability.

The Laptop Problem

Install on your daily machine and OpenClaw dies when you close the lid, lose WiFi, or restart for updates. Leads go cold. Automations break.

Dedicated Device

A Raspberry Pi or Mac Mini running 24/7 solves uptime — but you're responsible for power, networking, and security on your home connection.

VPS / Cloud (Best)

Datacenter-grade uptime, security isolation, and remote access from anywhere. Your AI agent runs independently of your personal devices.

RECOMMENDED FOR BUSINESS

How Do You Install OpenClaw on a VPS?

Deploy OpenClaw on a cloud server for 24/7 uptime, security isolation, and reliable business automation. Takes about 15 minutes.

Requirements

Ubuntu 22.04 or 24.04 (recommended)

2 GB RAM minimum (4 GB recommended)

10 GB+ disk space

Root or sudo access

Anthropic or OpenRouter API key

SSH client (Terminal, PuTTY, etc.)

Step 1: Provision Your VPS

Create an Ubuntu 22.04 or 24.04 VPS with at least 2 GB RAM. DigitalOcean, Hostinger, and Oracle Cloud (free tier) all work. If your VPS has less than 2 GB RAM, allocate a 4 GB swap file.

Step 2: SSH Into Your Server

Open your terminal and connect to your VPS as root. Replace YOUR_SERVER_IP with the IP from your provider's dashboard.

ssh root@YOUR_SERVER_IP

Step 3: Run the One-Line Installer

This script detects your OS, installs all dependencies (Node.js, Docker if needed), and launches the setup wizard. The --install-daemon flag registers OpenClaw as a background service so it survives reboots.

curl -sSL https://get.openclaw.ai/install.sh | bash --install-daemon

Step 4: Complete the Setup Wizard

The wizard walks you through: choosing your AI provider (Anthropic recommended), entering your API key, selecting messaging platforms to connect, and pairing your accounts. Choose "quick start" to skip advanced options.

Step 5: Verify Everything Works

Check the daemon is running, access the web dashboard, and test basic functionality. Save your OPENCLAW_GATEWAY_TOKEN — you'll need it for the dashboard.

openclaw status && openclaw test && openclaw logs --follow

Step 6: Harden Security

Bind the gateway to localhost (not 0.0.0.0), set up a firewall, and use Tailscale or an SSH tunnel for remote access. Never expose port 18789 to the public internet.

sudo ufw allow 22 && sudo ufw enable

Want Step-by-Step Guidance?

Our workshop walks you through setup, prompt configuration, and the 5 mistakes that trip up most people.

How Do You Install OpenClaw on a Local Device?

Install directly on your Mac, Linux machine, Windows PC, or Raspberry Pi. Great for experimenting and personal use.

Mac / Linux — One-Line Install

The fastest path. Detects your OS, installs dependencies, configures optimal settings, and starts the service.

curl -sSL https://get.openclaw.ai/install.sh | bash

Works on macOS (Apple Silicon), Ubuntu, Debian, and Raspberry Pi OS.

macOS — Homebrew

If you prefer Homebrew package management, install via the official tap.

brew install openclaw/tap/openclaw && openclaw init && openclaw start

Requires macOS 12+ with Homebrew installed.

Windows — WSL2

Install Ubuntu via WSL2, then run the standard Linux installer inside the WSL terminal.

wsl --install -d Ubuntu && wsl
curl -sSL https://get.openclaw.ai/install.sh | bash

Requires Windows 10/11. USB/serial hardware passthrough needs additional config.

Raspberry Pi (4 or 5)

Dedicated 24/7 hardware you control. The Pi 5 with 8 GB RAM is ideal — it can run local LLM models up to 7B parameters.

curl -sSL https://get.openclaw.ai/install.sh | bash --install-daemon

Invest in a USB-C UPS (~$25-40) and NVMe SSD via M.2 HAT instead of microSD to prevent corruption from power outages.

Docker (Any Platform)

Maximum isolation. Docker mounts ~/.openclaw for config and ~/openclaw/workspace as the agent's sandbox.

git clone https://github.com/openclaw/openclaw.git && cd openclaw && bash docker-setup.sh

Recommended for VPS deployments. Container isolation limits blast radius if something goes wrong.

Important: Local Install Limitations

✕OpenClaw stops when your device sleeps, loses power, or restarts — missed automations and cold leads.
✕Running on your primary machine means the AI has access to all your files, API tokens, and personal data.
✕Home internet outages take your agent offline — no follow-ups, no monitoring, no lead capture.
✕Security researchers found 135,000+ exposed OpenClaw instances from users who didn't configure firewalls.

For business automation, VPS deployment eliminates all of these issues. Our walks you through the entire VPS setup step by step.

Should You Install OpenClaw Locally or on a VPS?

Feature	Local Device	VPS / Cloud
24/7 Uptime	Only while powered on	Guaranteed — datacenter-grade
Security Isolation	Shares your personal machine	Isolated container — limited blast radius
Setup Complexity	One command	SSH + installer + firewall
Monthly Cost	$0 (your hardware)	$5-24/month
Power & Network	Depends on your home internet/power	Redundant datacenter infrastructure
Scalability	Limited by your hardware	Upgrade RAM/CPU in minutes
Remote Access	Requires port forwarding	Native — accessible from anywhere
Best For	Personal use, experimenting	Business automation, client work

VPS Cost Breakdown

Oracle Cloud

Free

4 ARM CPUs, 24 GB RAM

Hostinger VPS

$5-13/mo

1-Click Docker deploy

DigitalOcean

$6-24/mo

1-Click OpenClaw app

Any VPS + DIY

$5+/mo

15-min manual setup

What Should You Do After Installing OpenClaw?

After installing, connect your messaging platforms and verify everything works.

Verify the Daemon

Run openclaw status to check the service
Access dashboard at localhost:18789
Test with openclaw test
View logs: openclaw logs --follow

Connect Messaging

Telegram: Create bot via @BotFather → paste token
WhatsApp: Scan QR code or use Business API
Slack/Discord: Add app via OAuth
iMessage: Pair via the setup wizard

Configure Your Agent

Run openclaw configure for guided setup
Set user account and network preferences
Choose which skills to enable
Configure notification settings

Install Skills

Browse 10,700+ skills on ClawHub
Core integrations built-in
Thousands of community-built skills
Run openclaw skills list to see installed

How Do You Secure Your OpenClaw Installation?

OpenClaw is a high-privilege agent with shell access and API tokens. Treat it like a server, not a chatbot. Follow this checklist to stay safe.

Run on a dedicated device or VPS — never your primary work machine

Bind the gateway to 127.0.0.1, not 0.0.0.0

Use Tailscale for remote access (encrypted tunnel, invisible to internet)

Run inside Docker for container isolation

Enable command allowlisting (limit which terminal commands the AI can run)

Create a dedicated Linux user for OpenClaw with restricted permissions

Enable explicit consent mode: exec.ask: "on" in config

Treat ~/.openclaw like a password vault — it contains API keys

Run openclaw doctor after installation to audit your config

Never install community skills without reviewing the source code

For a deeper dive, read our complete OpenClaw Security Guide.

What If Something Goes Wrong During Installation?

"RPC probe: failed" error

Port 18789 is blocked or in use. Find the conflicting process: sudo lsof -i :18789 — then terminate it or change the OpenClaw port in config.

Gateway not responding or "0 tokens used"

The daemon stopped or API authentication failed. Run: openclaw gateway restart — then verify your API keys are correctly set in ~/.openclaw/config.

EACCES permission errors in Docker

The Node user (UID 1000) lacks permission to mounted volumes. Fix: sudo chown -R 1000:1000 ~/.openclaw ~/openclaw/workspace

Config files still say "clawdbot" not "openclaw"

Normal during the transition from ClawdBot to OpenClaw. Directories like ~/.clawdbot/ and CLAWDBOT_GATEWAY_TOKEN still work — the rename is cosmetic.

Installer fails on older Ubuntu

OpenClaw requires Node.js 22+. On older Ubuntu: curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash - && sudo apt-get install -y nodejs

WhatsApp/Telegram pairing not working

Ensure the gateway is running (openclaw status), your internet is connected, and your bot token is correct. For Telegram, make sure you created the bot via @BotFather and haven't revoked the token.

Your Competitors Are Already Automating. Are You?

Every week we send one automation that saves 10+ hours of manual work — the same playbooks our clients use to run their businesses on autopilot. Miss a week, miss the edge.

Save 10+ hours/week Cut AI costs by 97% Deploy in under 20 min

Get the Automation Playbook (Free)

One deploy-ready automation every week. Same strategies our clients pay thousands for. 400+ business owners already inside.

Need it done for you?

Book a Free Strategy Call See what we've built for real businesses →

Installation FAQ

Platform-Specific Guides

Mac / macOS Setup Windows / WSL2 Setup Linux / Ubuntu Setup Mac Mini Always-On Setup Docker Setup VPS / Cloud Deployment Troubleshooting Guide