Before You Begin

Getting moltbot up and running is straightforward. Discover why moltbot is the choice of 61,500+ developers and join the leading self-hosted AI assistant community. Here's what you'll need:

Prerequisites:
  • Operating System: macOS (recommended), or Linux/Windows via Docker
  • Node.js: Version 22 or higher (moltbot will check on first run)
  • AI Model Access: API key for Claude, GPT-4, or Ollama installed locally
  • 5 minutes: That's all the time you need for basic setup

Installation in 3 Steps

1 Download Moltbot

Head to the GitHub releases page and download the latest version:

  • macOS users: Download the .dmg installer (17.1 MB)
  • Other platforms: Download the .zip archive (21.3 MB)

2 Install

macOS: Double-click the .dmg file, drag moltbot to your Applications folder, and launch it.

ZIP archive: Extract the contents to your preferred location and run the moltbot executable from the terminal.

3 Initial Configuration

On first launch, moltbot's interactive setup wizard will guide you through:

  • AI Model Setup: Choose your AI provider (Claude, GPT-4, or local Ollama) and enter your API key
  • Gateway Configuration: The central hub starts automatically—no configuration needed
  • Channel Selection: Pick which messaging apps you want to connect (you can add more later)

Your First Automation

Once setup is complete, try sending moltbot a simple command through your connected messaging app:

"List all files in my Documents folder that were modified today"

Moltbot will execute the command and return the results. That's it! You've just automated your first task. Now try something more complex like managing emails, checking your calendar, or controlling smart home devices. See moltbot use cases for inspiration.

Explore 565+ Skills on ClawdHub

After your first automation, expand moltbot's capabilities with community-built skills from ClawdHub—the marketplace featuring 565+ skills following the AgentSkills standard:

npx clawdhub@latest install <skill-name>

Popular skills include GitHub integration, kubectl for Kubernetes, Home Assistant control, Notion sync, Trello management, and hundreds more. Each skill adds new automation capabilities without any coding required. Learn how moltbot's skills system works.

Next Steps: Connecting Channels

Moltbot's power comes from being accessible everywhere you already communicate:

  • WhatsApp: Run moltbot channel add whatsapp and scan the QR code
  • Telegram: Create a bot with @BotFather and add the token to moltbot
  • Discord/Slack: Follow the interactive setup wizard for OAuth integration
  • iMessage: Available on macOS with native integration

Each channel takes just a few minutes to configure. Detailed instructions for every platform are in the comprehensive documentation.

Common Setup Questions

Do I need to keep moltbot running 24/7?
Only if you want to receive messages and run automation anytime. For personal use, you can run moltbot when needed. For always-on access, deploy it to a server or Raspberry Pi.
Which AI model should I choose?
Claude (Anthropic) is recommended for best performance with moltbot. GPT-4 is also excellent. For complete privacy, use Ollama with local models—it's slower but keeps everything offline.
Can I use moltbot without an API key?
Yes, by installing Ollama and using local models. This is completely free but requires more powerful hardware and is slower than cloud AI models.
What if I get stuck during setup?
Join the Discord community for real-time help from the community, or check the documentation for troubleshooting guides. Also review our moltbot security guide for best practices.
Download Moltbot See Use Cases