Clawdbot (OpenClaw): Complete Setup Guide + Security Concerns You Need to Know

OpenClaw is an open-source, self-hosted personal AI assistant created by Peter Steinberger (@steipete), founder of PSPDFKit.

Unlike ChatGPT or Claude that live in browser tabs, OpenClaw:

• Runs on YOUR hardware - your data stays local
• Integrates with messaging apps - WhatsApp, Telegram, Discord, Slack, Signal, iMessage
• Actually does things - not just chat, but execute actions

Think of it as "Claude with hands."

The numbers are staggering:

• 9,000 GitHub stars in the first 24 hours
• 80,000+ stars within days
• 100,000+ stars within two months
• One of the fastest-growing open-source projects ever

The appeal? People are tired of:

1. Sending their data to cloud AI services
2. AI that can only talk, not act
3. Being locked into single platforms

OpenClaw solves all three.

Multi-Platform Messaging

Works with channels you already use:

• WhatsApp
• Telegram
• Discord
• Slack
• Signal
• iMessage
• Microsoft Teams
• Google Chat

50+ Service Integrations

• Email: Gmail, Outlook
• Calendar: Google Calendar
• Tasks: Todoist, Notion
• Smart Home: Philips Hue
• Dev Tools: GitHub, terminal commands
• Creative: ElevenLabs, image generation
• Fitness: Health tracking

Privacy-First Architecture

• Self-hosted on your own devices
• No data sent to third-party servers
• Works with local LLMs via Docker Model Runner
• Complete control over your information

On January 27, 2026, Anthropic (makers of Claude) sent a trademark request. The original name "Clawdbot" was too close to "Claude."

Steinberger's response was characteristically upbeat: "Anthropic asked us to change our name (trademark stuff), and honestly? 'Molt' fits perfectly—it's what lobsters do to grow."

The project went through:

1. Clawdbot (original)
2. Moltbot (brief transition)
3. OpenClaw (current)

Prerequisites

• Docker installed
• A machine that can run 24/7 (old laptop, Raspberry Pi, NAS, etc.)
• API keys for LLM provider (or run local models)

Quick Setup

# Clone the repository
git clone https://github.com/clawdbot/clawdbot.git
cd clawdbot
 
# Configure your channels
cp .env.example .env
# Edit .env with your API keys and channel configs
 
# Run with Docker
docker-compose up -d

Connect Your First Channel

1. Start with Telegram (easiest setup)
2. Create a bot via @BotFather
3. Add the token to your config
4. Message your bot - OpenClaw responds!

The killer feature for privacy advocates: run everything locally.

# Use Docker Model Runner for local LLMs
docker model pull llama3.2
 
# Configure OpenClaw to use local model
# In .env:
LLM_PROVIDER=docker-model-runner
LLM_MODEL=llama3.2

Now your AI assistant never sends data anywhere.

Real examples from users:

• "Check me in for my flight tomorrow" ✅
• "Turn off the living room lights" ✅
• "Summarize my unread emails" ✅
• "Create a GitHub issue for this bug" ✅
• "Add milk to my shopping list" ✅
• "What's on my calendar this week?" ✅

It's not just answering questions - it's taking actions.

OpenClaw represents a trend: AI moving from cloud to edge.

Why this matters:

1. Privacy: Your data, your control
2. Customization: Modify the code, add integrations
3. Reliability: Works offline, no API outages
4. Cost: No per-token charges after setup

We're seeing the beginning of truly personal AI - assistants that know everything about you because you trust them with that data, running on hardware you own.

Yes, if you:

• Value privacy and data ownership
• Are comfortable with Docker/self-hosting
• Want an AI that integrates with your life
• Like tinkering with open-source projects

Maybe wait if you:

• Need plug-and-play simplicity
• Don't have hardware to run it 24/7
• Are happy with cloud AI assistants

• Official Website
• GitHub Repository
• Docker Model Runner Integration Source: hereshecodes.app by Danielle Scantling