Moltbot Complete Guide: Build Your Own 24/7 AI Assistant on Mac Mini

Imagine having your own personal JARVIS—an AI assistant that runs 24/7, remembers everything about you, manages your emails, schedules your meetings, controls your computer, and proactively updates you on important events. This is no longer science fiction. Welcome to Moltbot (formerly known as Clawdbot), the open-source, self-hosted AI assistant that's causing developers to rush out and buy Mac Minis specifically to run their own personal AI agents.

In this comprehensive guide, we'll cover everything from installation to advanced features, explore why the Mac Mini has become the go-to hardware for AI enthusiasts, and show you real-world examples of what Moltbot can do.

What is Moltbot?

Moltbot is an open-source, self-hosted AI assistant created by Peter Steinberger. Unlike cloud-based AI assistants like Siri or Alexa, Moltbot runs entirely on your own machine. Your data, your conversations, your context, and your skills all stay on your hard drive—giving you complete privacy and control.

The project was originally called Clawdbot (a playful reference to Anthropic's Claude AI with a lobster twist), but was renamed to Moltbot in early 2026 after a trademark-related discussion. The mascot? A charming space lobster named Molty. 🦞

Key Differentiators

Feature	Traditional AI Assistants	Moltbot
Data Privacy	Sent to cloud servers	Stays on your machine
Customization	Limited	Fully extensible with skills
Proactive Actions	Mostly reactive	Can initiate tasks autonomously
System Access	Sandboxed	Full computer control (optional)
Memory	Session-based	Persistent across all conversations
Platforms	Single ecosystem	Works on WhatsApp, Telegram, Discord, Signal, iMessage

Why Are Developers Buying Mac Minis for This?

The Mac Mini M4 has become the hardware of choice for running Moltbot, and there's been a noticeable surge in Mac Mini purchases specifically for this purpose. Here's why:

1. Incredibly Low Power Consumption

The Mac Mini with Apple Silicon can idle at just 5 watts and operate under load at around 20 watts. Running 24/7, this translates to potentially less than $5 per year in electricity costs. Compare that to a traditional desktop PC, which might consume 100-200 watts.

2. Silent Operation

The Mac Mini runs virtually silent, making it perfect for a home office or living room setup. You can leave it running in the corner without any fan noise disturbing your environment.

3. Compact Footprint

At just 7.7 inches square and 1.4 inches tall, the Mac Mini takes up minimal desk space. Many users tuck it behind their monitor or in a drawer.

4. Powerful Apple Silicon

The M4 chip features a 10-core CPU, 10-core GPU, and a Neural Engine specifically designed for machine learning inference. This means local AI operations run blazingly fast.

5. Automatic Restart on Power Failure

The Mac Mini can be configured in System Preferences to automatically start up after a power failure—perfect for an always-on AI server.

6. Affordable Entry Point

Starting at $599 for the base M4 model (or 15% less refurbished), it's an accessible entry point compared to building a dedicated AI server.

"It feels like we're living in the future." — @Philo01 on X (formerly Twitter)

Moltbot Features Deep Dive

🖥️ Runs on Your Machine

Moltbot runs locally on Mac, Windows (via WSL2), or Linux. It supports multiple AI providers:

Anthropic Claude (recommended)
OpenAI GPT-4/GPT-4o
Local models via Ollama, LM Studio, or similar

All settings, memories, and commands are stored directly on your hard drive.

💬 Any Chat App Integration

Talk to your AI assistant through the apps you already use:

WhatsApp (QR code login)
Telegram (bot token)
Discord (bot integration)
Slack
Signal
iMessage
Mattermost

This means you can message your AI from your phone while you're out, and it will execute tasks on your home computer.

🧠 Persistent Memory

Unlike ChatGPT or Claude web interfaces that forget everything between sessions, Moltbot remembers:

Your preferences and communication style
Previous projects and conversations
Important dates and recurring tasks
Your work context and relationships

Over time, it becomes uniquely yours—understanding your patterns and anticipating your needs.

🌐 Browser Control

Moltbot can:

Navigate to any website
Fill out forms automatically
Extract data from web pages
Take screenshots
Interact with web applications

This enables automation of tasks like:

Checking flight status
Price monitoring
Form submissions
Web scraping for research

💻 Full System Access

With your permission, Moltbot can:

Read and write files anywhere on your system
Execute terminal commands and scripts
Manage applications (open, close, interact)
Control the shell for development tasks
Run scheduled cron jobs for automation

You can configure sandbox modes for security:

"off" — Full access (for power users)
"non-main" — Sandboxed for non-primary agents
Read-only modes for sensitive directories

🔌 Skills & Plugins

Moltbot is extensible through a skills system:

Install community-created skills from ClawdHub
Build your own custom skills
The AI can even write its own skills to extend its capabilities

Popular skills include:

Email management and triage
Calendar scheduling
Home automation (Homebridge, Home Assistant)
Stock portfolio monitoring
News and RSS aggregation

Complete Installation Guide

Prerequisites

Before installing Moltbot, ensure you have:

Node.js version 22 or higher
A computer running macOS, Linux, or Windows (WSL2)
An API key for your preferred AI provider (Anthropic Claude recommended)
Optional: Brave Search API key for web search functionality

Step 1: Install the CLI

The fastest way to install Moltbot is via the one-liner script:

macOS / Linux:

curl -fsSL https://molt.bot/install.sh | bash

Windows (PowerShell):

iwr -useb https://molt.bot/install.ps1 | iex

Alternative: Install via npm:

npm install -g moltbot@latest

Or with pnpm:

pnpm add -g moltbot@latest

The installer will automatically set up Node.js and all dependencies for you.

Step 2: Run the Onboarding Wizard

After installation, run the onboarding wizard to configure everything:

moltbot onboard --install-daemon

The wizard will guide you through:

Gateway Configuration — Local vs remote gateway setup
Authentication — Setting up your AI provider (Anthropic API key, OpenAI, or OAuth)
Chat Providers — Connecting WhatsApp, Telegram, Discord, etc.
Daemon Installation — Running as a background service (launchd on macOS, systemd on Linux)

Step 3: Configure Authentication

For the best experience, we recommend using an Anthropic API key. The wizard can store it securely for service use.

Authentication files are stored in:

OAuth credentials: ~/.clawdbot/credentials/oauth.json
Auth profiles: ~/.clawdbot/agents/<agentId>/agent/auth-profiles.json

You can also use Claude Code credentials if you have a Claude Pro subscription:

claude setup-token

Step 4: Configure Web Search (Optional but Recommended)

To enable web search capabilities, set up a Brave Search API key:

moltbot configure --section web

This stores your API key in tools.web.search.apiKey.

Step 5: Start the Gateway

Check if the gateway is already running:

moltbot gateway status

If not running, start it:

moltbot gateway --port 18789 --verbose

The gateway will be accessible at http://127.0.0.1:18789/

Step 6: Quick Verification

Verify your installation is working correctly:

moltbot status
moltbot health
moltbot security audit --deep

All checks should pass before proceeding.

Step 7: Connect Your First Chat Platform

WhatsApp (QR Login)

WhatsApp is the most popular integration. Connect using QR code login:

moltbot channels login

A QR code will appear in your terminal. Scan it with your WhatsApp app (Settings → Linked Devices → Link a Device).

Create a bot via @BotFather on Telegram, get your bot token, and configure it in Moltbot.

Discord

Create a Discord application, get your bot token, and add it to your server with appropriate permissions.

Step 8: DM Safety and Pairing Approvals

For security, Moltbot uses a pairing approval system. New contacts must be approved before they can interact with your bot:

# List pending pairing requests
moltbot pairing list whatsapp

# Approve a specific request
moltbot pairing approve whatsapp <code>

Step 9: Verify End-to-End

Send a test message to verify everything works:

moltbot message send --target +15555550123 --message "Hello from Moltbot"

Or simply message yourself on WhatsApp and watch the magic happen!

Building from Source (For Developers)

If you want to contribute or customize Moltbot:

# Clone the repository
git clone https://github.com/moltbot/moltbot.git

# Navigate to the directory
cd moltbot

# Install dependencies
pnpm install

# Build the UI
pnpm ui:build

# Build the project
pnpm build

# Run the onboarding wizard
moltbot onboard --install-daemon

For development, you can run the gateway directly:

node moltbot.mjs gateway --port 18789 --verbose

Web Dashboard

Moltbot includes a beautiful web dashboard for monitoring and configuration:

moltbot dashboard

Access it at http://127.0.0.1:18789/

The dashboard provides:

Real-time conversation monitoring
Skill management
Agent configuration
Security settings
Health monitoring

Real-World Use Cases

📧 Email Management

"Summarize my unread emails and highlight anything urgent."

Moltbot can connect to your email, summarize newsletters, flag important messages, and even draft replies for your approval.

📅 Calendar & Scheduling

"Schedule a meeting with John next week when we're both free."

With calendar integration, Moltbot can check availability, send invitations, and manage your schedule proactively.

✈️ Flight Check-In

"Check me in for my Southwest flight tomorrow at the earliest time."

Using browser automation, Moltbot can log into airline websites and complete check-ins automatically.

🏠 Home Automation

"Turn off all the lights and set the thermostat to 68°F."

With Home Assistant or Homebridge integration, control your smart home via natural conversation.

📊 Research Assistant

"Find the latest news about Apple and summarize the key points."

Web search combined with AI summarization makes research effortless.

💻 Development Automation

"Run the test suite and fix any failing tests."

For developers, Moltbot can execute shell commands, run builds, and even commit code changes.

⏰ Proactive Updates

Unlike passive assistants, Moltbot can:

Send you a morning briefing with weather, calendar, and news
Alert you when a package is delivered
Remind you of tasks at the right time
Monitor RSS feeds and social media for keywords

Security Best Practices

With great power comes great responsibility. Moltbot has extensive system access, so follow these security best practices:

1. Use Sandbox Modes

For sensitive environments, configure sandbox mode in your settings:

{
  "routing": {
    "agents": {
      "main": {
        "workspace": "~/clawd",
        "sandbox": {
          "mode": "non-main"
        }
      }
    }
  }
}

2. Use Dedicated Email Accounts

Create a separate email account for Moltbot rather than granting access to your primary account.

3. Network Sandboxing

Consider running Moltbot in a network-isolated environment if you're concerned about external access.

4. Regular Security Audits

Run deep security audits regularly:

moltbot security audit --deep

5. Monitor Activity Logs

Review Moltbot's activity logs to ensure it's only performing expected actions.

6. Pairing Approvals

Always use the pairing approval system to control who can interact with your bot.

Cost Breakdown

Moltbot itself is free and open-source. However, running it involves some costs:

Item	Cost
Moltbot Software	Free
Mac Mini M4	$599 one-time
Anthropic API	~$20-50/month depending on usage
Electricity	~$5/year
Brave Search API	Free tier available

Total Monthly Operating Cost: ~$25-50/month for typical usage

This is remarkably affordable for what essentially becomes a 24/7 AI employee handling your mundane tasks.

Community and Resources

Moltbot has a thriving community:

GitHub: 30,000+ stars, 130+ contributors
Discord: 9,000+ members
ClawdHub: Marketplace for community skills

Helpful Links

Homepage: clawd.bot
Documentation: docs.molt.bot
GitHub Repository: github.com/moltbot/moltbot
Skills Marketplace: clawdhub.com
Discord Community: discord.com/invite/clawd

Troubleshooting Common Issues

Gateway Won't Start

Check if another process is using port 18789:

lsof -i :18789

Kill any conflicting processes or use a different port.

WhatsApp QR Code Not Appearing

Ensure you have the latest version installed:

npm update -g moltbot

API Key Issues

Verify your API key is correctly stored:

moltbot configure --section auth

High Memory Usage

If memory usage is high, consider adjusting the default model or reducing memory retention settings.

The Future of Personal AI

Moltbot represents a fundamental shift in how we interact with AI. Instead of typing queries into a website and receiving text responses, we now have AI agents that:

Take action on our behalf
Remember our context over time
Integrate with our existing tools
Run privately on our own hardware

The fact that developers are specifically buying Mac Minis to run their own AI agents shows that the era of personal AI infrastructure is here. We're not just using AI—we're hosting AI.

Whether you're a developer looking to automate your workflow, a business professional wanting to optimize your time, or an enthusiast excited about the future of personal computing, Moltbot offers a glimpse into what's possible when you have your own AI assistant running 24/7.

Conclusion

Setting up Moltbot is straightforward:

Install via the one-liner script
Run the onboarding wizard
Connect your favorite chat platforms
Start talking to your AI

Within 30 minutes, you can have your own personal AI assistant running on your Mac Mini, accessible from WhatsApp, and ready to manage your digital life.

The future of personal AI isn't in the cloud—it's on your desk.

Happy automating! 🦞

Frequently Asked Questions

Is Moltbot free?

Yes, Moltbot is completely free and open-source. You only pay for the AI API usage (e.g., Anthropic, OpenAI) and your hardware.

Does it work on Windows?

Yes, via Windows Subsystem for Linux 2 (WSL2). Native Windows support is in development.

Can I use local language models?

Yes, Moltbot supports local models via Ollama, LM Studio, and similar tools. However, for the best experience, cloud APIs like Claude are recommended.

Is it safe to give an AI full computer access?

It comes with risks. Use sandbox modes, dedicated accounts, and follow security best practices. The open-source nature means the community actively reviews the code for vulnerabilities.

Can multiple people use the same Moltbot instance?

Yes, through group chats and multiple paired accounts. Each person can be managed through the pairing system.

What's the difference between Clawdbot and Moltbot?

They're the same project. Clawdbot was renamed to Moltbot in early 2026. The mascot lobster's name is Molty.