Two ways to run your personal AI agent — pick the one that fits your setup
Turn what you learned into a concrete stack decision.
01AI Agent ToolboxBuild autonomous AI agents that can read, write, and actWant the shortlist in your inbox?
Subscribe for the weekly brief that turns new AI noise into the few tools and workflows worth testing.
Curated bundles that help you move from this guide into a working stack.
Guide
OpenClaw Infrastructure: Dedicated Machine vs. Cloud VPS (Advanced)
LaunchAgents, PM2, systemd, SSH tunnels — build a production-grade OpenClaw deployment that survives reboots and network drops
Guide
Installing OpenClaw: Zero to Running Agent
Get your own AI agent with personality, memory, and Telegram access in under 30 minutes
Guide
Claude Cowork + Claude Agents: Your First Setup
How to use Claude as a real AI coworker — not just a chat window — with agents that take action on your behalf
When you're setting up OpenClaw, the first decision isn't which model to use or what skills to add — it's where OpenClaw actually lives.
You have two options:
Neither is wrong. They just suit different situations. This guide walks you through both so you can pick what fits.
This is the gold standard for anyone serious about having an AI agent that's always available.
A second computer — usually a Mac mini, an old MacBook, or a small Windows PC — that you leave plugged in and connected to the internet at all times. OpenClaw runs on it in the background. You never have to "start" it — it's just always there.
1. Install OpenClaw
npm install -g openclaw
Or with Bun (faster):
bun install -g openclaw
2. Run first-time setup
openclaw init
This walks you through naming your agent, connecting a model provider, and linking your Telegram bot.
3. Set it up as a background service (macOS)
So OpenClaw starts automatically when the machine boots:
openclaw service install
Then start it:
openclaw service start
4. Verify it's running
Open Telegram, message your bot, and say "hello." If it replies, you're live.
If you don't have a spare computer, no problem. You have two paths: run OpenClaw on your main laptop, or rent a cloud server.
This works fine for getting started. The tradeoff is that OpenClaw only runs when your laptop is on and awake.
Best for: Trying OpenClaw out, casual use, one-off tasks
How to run it:
openclaw start
Leave a terminal window open. When you're done for the day, close it. Your agent won't be active while your laptop is sleeping — but all your settings, memory, and configuration persist.
Practical tip: Don't run local AI models here unless you have a powerful laptop. Use a cloud model provider (like Anthropic Claude API or OpenAI) instead — they're faster and don't drain your battery.
This gives you the "always-on" benefit of a dedicated machine without needing to own extra hardware. You rent a small Linux server online and run OpenClaw there.
Best for: Always-on without owning extra hardware, running 24/7 tasks, teams or multiple projects
Recommended providers (cheapest to most powerful):
| Provider | Monthly cost | Good for | |---|---|---| | Hetzner | ~$5/mo | Solo builders, lean setups | | DigitalOcean | ~$6/mo | Reliable, great docs for beginners | | Vultr | ~$6/mo | Good global locations | | Railway | Pay-per-use | Easy deploy, no server management |
High-level setup on a VPS:
npm install -g openclawopenclaw init and complete setup# Install pm2 (process manager)
npm install -g pm2
# Start OpenClaw with pm2
pm2 start openclaw -- start
pm2 save
pm2 startup
Cloud model note: On a VPS, always use cloud model providers (Anthropic API, OpenAI, etc.) — not local models. VPS hardware can't run local models well.
| Factor | Dedicated machine | Daily laptop | Cloud VPS | |---|---|---|---| | Always-on | Yes | No | Yes | | Upfront cost | Hardware cost | $0 | $0 | | Monthly cost | Electricity | $0 | $5–20/mo | | Local AI models | Yes (with enough RAM) | Only if powerful | No | | Setup difficulty | Easy | Easiest | Moderate | | Best for | Power users | Getting started | Always-on without hardware |
If you're just getting started: Run on your daily laptop. Get familiar with OpenClaw first, then upgrade your setup.
If you want always-on from day one: Cloud VPS is the fastest path — you can have it running in under an hour.
If you're building seriously: A dedicated Mac mini is the best long-term setup. One-time cost, zero ongoing friction.
Want hands-on help choosing the right setup or getting OpenClaw running? Book a free 30-minute call and we'll walk through it together.