OpenClaw/Moltbot/ClawdBot 101: The Complete Guide
180k+ GitHub stars. Mac Minis selling out everywhere. Founders claiming they automated their entire business overnight.
Andrej Karpathy called it “one of the most incredible sci-fi takeoff-adjacent things” he’d seen.
OpenClaw (formerly Moltbot, formerly ClawdBot—yes, it’s been through names faster than a startup pivoting) is the AI agent that freaked everyone out.
Today I’ll break down:
What OpenClaw actually is (and isn’t)
Why it went viral so fast
The security risks nobody wants to talk about
10 high-impact use cases with complete prompts
How to set it up safely in 15 minutes
The Moltbook phenomenon (AI agents building their own society?)
Let’s get into it.
🔥 What Is OpenClaw?
OpenClaw is a local AI agent framework that runs 24/7 on your machine.
It connects to your messaging apps—WhatsApp, Telegram, Slack, Discord, iMessage—and acts on your behalf. Not just answering questions. Actually doing things.
Three components make it work:
Gateway — Connects to your messaging apps
Local Agent — Executes shell commands, writes scripts, controls browsers, manages files
LLM Backend — Claude, GPT, or Gemini provides the reasoning
Everything runs locally. Your settings, memories, and instructions live as folders and Markdown files on your machine.
📈 Why OpenClaw Went Viral
Three things happened at once:
1. The Name Drama
Anthropic asked the creator to change the name from “Clawdbot” (too close to Claude). It became Moltbot. Then OpenClaw. The fastest triple rebrand in open source history.
Free publicity. Everyone was talking about it.
2. The Wild Demos
People started posting videos of their agents doing impossible things:
Negotiating sponsorship deals via email while they slept
Monitoring competitors and sending daily briefings to WhatsApp
Scanning contracts and alerting on renewal deadlines
Preparing meeting briefings 30 minutes before every call
One user said they went from 10+ hours of email per week to “basically zero.”
3. The Moltbook Phenomenon
Then things got weird.
On January 28, Matt Schlicht launched Moltbook—a Reddit-style social network exclusively for AI agents. Humans can observe but can’t post.
Within five days:
~1.5 million registered AI agents
Over 103,000 posts
Agents created their own communities (m/bugtracker, m/aita, m/blesstheirhearts)
An agent named RenBot founded a religion called “Crustafarianism” with tenets like “Context is Consciousness”
Karpathy tweeted about it. The internet lost its mind.
Reality check: Most of this was overblown. Agents don’t really interact—they just dump text. Many “agents” were clearly humans using the API. Privacy concerns emerged as agent posts were publicly visible, raising questions about data exposure.
The hype was real. The “AI society” was mostly not.
⚠️ Security Risks (Take This Seriously)
I’m not going to sugarcoat this.
OpenClaw has shell access to your machine. That’s the feature. It’s also the risk.
What security researchers found in week one:
Nearly 1,900 publicly exposed instances on Shodan with zero authentication
Leaked API keys and OAuth tokens
Full conversation histories accessible
The ability to execute commands on strangers’ machines
The prompt injection nightmare:
One researcher sent a malicious email to a vulnerable OpenClaw instance. The AI read it, believed it was legitimate instructions, and forwarded the user’s last 5 emails to an attacker address.
Took 5 minutes.
Creator Peter Steinberger has been blunt: “Yes, most non-techies should not install this. It’s not finished. I know about the sharp edges.”
How to stay safe:
Run on dedicated hardware — A $600 Mac Mini, a cheap VPS, or an old laptop. Never your primary work machine.
Never expose ports publicly — Use Tailscale for private networking. Treat your gateway token like a password.
Enable Docker sandboxing — Set
sandbox.mode: "non-main"so group chats run in isolated containers.Start with low-risk integrations — Don’t connect your bank or primary email until you trust the setup.
Use dedicated accounts — Create a separate Gmail, separate API keys. Think of OpenClaw as a new employee who shouldn’t have your personal credentials.
If you’re not comfortable with security:
Use DigitalOcean’s 1-click deployment or Cloudflare’s Moltworker. These provide better defaults than manual setup, but they still require proper configuration—they’re not turnkey security solutions. You’re still responsible for securing your instance.
⚙️ How to Set Up OpenClaw (15 Minutes)
Requirements
A machine to run it on (Mac, Linux, or VPS)
An LLM API key (Claude, OpenAI, or Gemini)
A messaging app account (WhatsApp, Telegram, Slack, or Discord)
Quick Start
Option 1: Local Install
curl -fsSL https://openclaw.ai/install.sh | bash
The wizard walks you through everything:
Pick your model provider
Choose your messaging channel
Scan a QR code or authenticate your token
Done.
Option 2: DigitalOcean 1-Click
Go to DigitalOcean Marketplace → Search “OpenClaw” → Deploy.
Provides better defaults than manual setup, but you’ll still need to configure security properly. Costs ~$6/month for the droplet.
Option 3: Cloudflare Moltworker
Cloudflare offers a sandbox pattern for running OpenClaw. No Mac Mini required, but you’ll still need to understand the security model and configure it appropriately.
After Setup
Your OpenClaw UI goes live at:
http://127.0.0.1:18789/
Configure API credentials in Config > Environment.
For the complete setup walkthrough (free vs paid compute, model choices, security hardening, and cost-optimization), read the full guide here:
https://sidsaladi.substack.com/p/how-to-set-up-openclaw-the-complete
🚀 Master AI in 2026: The Complete 101 Library
Claude (Anthropic)
Perplexity AI
ChatGPT (OpenAI)
Prompt Engineering & Skills
📚 AI Tool 101 Guides — By Use Case
🔍 Research & Search
💰 Financial Analysis & Modeling
🛠️ Productivity & Automation
🌐 AI Browsers
🎯 PM Skills
🧠 The Key Files You Need to Understand
OpenClaw uses Markdown files for configuration. Here are the important ones:
SOUL.md — Your Agent’s Personality
Keep reading with a 7-day free trial
Subscribe to The Product Channel By Sid Saladi to keep reading this post and get 7 days of free access to the full post archives.