Do I need expensive hardware to run OpenClaw?

No. OpenClaw runs on a free AWS EC2 instance with no hardware purchase, or on a mini PC starting around $150-300. The Ryzen 7 mini PC used throughout these tutorials runs both Ollama and OpenClaw simultaneously without issue.

What is the difference between the AWS and mini PC setup?

AWS EC2 is free to start, runs 24/7 in the cloud, and requires zero hardware purchase - great for testing the concept first. A mini PC is a one-time cost ($300-500) that pays for itself within a year, and your data never leaves your home network.

Your Self-Hosted AI
Sanctuary

Q: What is OpenClaw?

OpenClaw is a self-hosted AI agent you install on your own hardware - an AWS EC2 instance, a mini PC, or a home server. It connects to Discord and can use local LLMs via Ollama or cloud AI APIs. You own it completely. No subscriptions required.

Q: Is OpenClaw free to use?

Yes. OpenClaw is free and open source. You pay for your hardware or cloud instance - that is it. No monthly fees, no per-message billing, no vendor lock-in. A mini PC pays for itself within months compared to cloud AI subscriptions.

Q: Can OpenClaw run AI without an internet connection?

Yes, using Ollama. Install Ollama on your local hardware, pull a model like Qwen3:8b or Llama 3.2, and OpenClaw runs entirely offline. No API calls, no data leaving your machine, no usage caps. Book III covers the whole setup.

Build it. Own it. Never rent again.

OpenClaw is a personal AI agent that runs on hardware you control - a cloud server, a mini PC, a home server in your closet. Hook it up to Discord. Give it a local brain with Ollama. Your data stays yours. No subscriptions to cancel. No monthly bills to the big boys.

These guides walk you through every step. Start where you are, go at your own pace, and build something that's genuinely yours.

Book I

Get It Running

Pick your platform. Both paths lead to the same place - a running, Discord-connected OpenClaw agent you own completely. The cloud path costs nothing upfront. The local path costs once and pays for itself within the year.

AWS EC2 Setup ●○○Free-tier cloud instance, SSH hardened, Discord connected. Already know Linux? There's a quick-start version inside. Mini PC Setup ●○○Your hardware, no monthly bills, data never leaves your home

Book II

Fortify Your Sanctuary

A running agent is just the beginning. These guides turn a generic AI into something personal - one that knows who you are, protects your credentials, follows your rules, and gets sharper over time.

Memory Architecture ●○○Build an agent with real context, real preferences, and real rules Securing Your Secrets ●○○Move API keys out of plaintext config files before you build skills that use them Skills & Slash Commands ●●○Teach your agent what to do with plain markdown files Hooks ●●○Protect sensitive files, block deploys until tests pass, stay in control Telegram Bot Setup NEW ●○○Prefer Telegram over Discord? Swap it in - BotFather token, one config line

Book III

Build Your Local AI Stack

Stop renting intelligence. Run real language models on your own hardware with no API fees and no data leaving your machine. Start with Ollama Basics and Ollama Advanced - everything else is optional from there.

Ollama Basics CORE ●○○Install Ollama and run your first local model - no API keys, no internet required Best Local AI Models 2026 ●○○Qwen3 vs Llama vs DeepSeek vs Phi4 - find the right model for your hardware Ollama Advanced CORE ●●○Tune parameters and wire Ollama into OpenClaw - this is where your bot gets a local brain Open WebUI ●○○Your own private ChatGPT interface - conversation history, document upload, multi-user Document Q&A with RAG ●●○Ask questions against your own PDFs and notes, all local Local Coding Assistant ●○○Tab autocomplete in VS Code powered by Ollama - your code never leaves the machine Run Gemma 4 Locally ●○○The only local model with native image input - analyze photos, read screenshots Ollama + MCP NEW ●●○Give your local AI real tools - read files, fetch live web content, query databases Securing Your Local AI NEW ●●●Lock it down so it is invisible from the internet but reachable from anywhere you are

The Context Window

Short takes on local AI - new models, tools, and what's actually worth your attention.

Jun 26, 2026 · Models

Gemma 4 Vision: What It Can Actually Do on Your Local Machine

Gemma 4 is the only local model small enough for consumer hardware that accepts image input. Here's what it's genuinely good at, what it struggles with, and how to use it.

Read →

Jun 26, 2026 · Ollama

Qwen3's Thinking Mode: When to Use It and How to Turn It Off

qwen3:8b reasons through problems before answering - useful for complex tasks, wasteful for simple ones. Here's what it is, how to toggle it per prompt, and how it eats your context window.

Read →

Common Questions

What is OpenClaw?

A self-hosted AI agent you run on your own hardware - an AWS EC2 instance, a mini PC, or a home server. It connects to Discord and can use local LLMs via Ollama or cloud AI APIs. You own it completely. No subscriptions.

Do I need expensive hardware?

No. It runs on a free AWS EC2 instance (zero upfront cost) or a mini PC starting around $150. Most of these tutorials were built on a $479 Ryzen 7 mini PC that runs Ollama and OpenClaw simultaneously without breaking a sweat.

Is it actually free?

The software is free and open source. You pay for your hardware or cloud time - that's it. No monthly fees, no per-message billing, no vendor lock-in. A mini PC pays for itself within months vs. cloud AI subscriptions.

Can it run without an internet connection?

Yes. Install Ollama, pull a model like Qwen3:8b or Llama 3.2, and OpenClaw runs entirely offline. No API calls, no data leaving your machine, no usage caps. Book III covers the whole setup.

AWS or mini PC - which should I pick?

AWS is free to start and runs 24/7 with no hardware purchase. Great for proving the concept first. A mini PC is a one-time cost that pays for itself within months, and your data never leaves your home. Both paths lead to the same place.

Something not working?

The troubleshooting guide covers the most common errors - Discord bot offline, Ollama not loading, gateway won't start - with actual commands to fix them.