
OpenClaw (ex. Clawdbot / Moltbot): Your Personal AI Assistant on Your Own Device
Discover OpenClaw, the personal AI assistant that runs on your own device. Stay in control of your data, connect to WhatsApp, Telegram, Discord, and more, and enhance productivity with secure, extensible AI tools. Learn setup, features, and advanced usage.
In recent years, AI assistants have become essential tools for both businesses and individuals to significantly improve productivity. However, most AI assistants operate in the cloud, storing user data on external servers. This can raise concerns about privacy and security.
Enter OpenClaw - a fully personal AI assistant that runs on your own device and integrates directly with the chat apps you already use.
Interestingly, this project has gone through several name changes. It started as Clawd, a playful nod to Anthropic’s Claude. Legal considerations led to a change, and the name became Moltbot, symbolizing growth through “molting” like a lobster shedding its shell. Finally, the project settled on OpenClaw, combining the ideas of open source and its lobster-inspired heritage.
In this article, we’ll cover OpenClaw’s features, setup process, usage, and security model — everything you need to get started from scratch.
Why OpenClaw Stands Out
OpenClaw differs from traditional cloud-based AI assistants in several key ways:
Runs Locally on Your Device
OpenClaw can run on a laptop, home server, or VPS. You control the AI on your own infrastructure, keeping data and keys fully under your management.
Multi-Channel Support
OpenClaw works with WhatsApp, Telegram, Discord, Slack, Google Chat, Signal, iMessage, Microsoft Teams, and more. Additional channels can be easily added.
Flexible Security Design
By default, unknown DMs are not processed until explicitly approved via a pairing code. Group and channel sessions can also be sandboxed individually.
Extensible Skill Platform
Through ClawdHub, you can add new skills or automatically update them. Developers can easily create and register custom skills.
Voice and Visual Interaction
On macOS, iOS, and Android, OpenClaw supports Voice Wake, Talk Mode, and Canvas, a visual workspace for agent-driven interactions, enabling natural conversation and screen operations.
Getting Started with OpenClaw
Here’s a step-by-step guide to set up OpenClaw from zero to first working chat.
1. Check Prerequisites
Ensure your environment is ready:
- Node.js ≥22
- pnpm (optional, recommended for building from source)
- macOS: Xcode/CLT if building apps
- Windows: WSL2 (Ubuntu recommended)
- Brave Search API key (optional, for web search functionality)
2. Install the CLI
OpenClaw is installed via CLI:
macOS / Linux:
Codecurl -fsSL https://openclaw.bot/install.sh | bash
Windows (PowerShell):
Codeiwr -useb https://openclaw.ai/install.ps1 | iex
Alternative global install:
Codenpm install -g openclaw@latest pnpm add -g openclaw@latest
3. Run the Onboarding Wizard
Use the CLI wizard to safely set up OpenClaw:
Codeopenclaw onboard --install-daemon
The wizard sets up:
- AI model and authentication (OAuth recommended)
- Gateway settings (local or remote)
- Channels (WhatsApp, Telegram, Discord, Mattermost, etc.)
- DM pairing settings for secure messaging
- Workspace and initial skills
- Optional background service installation
This provides a ready-to-use AI assistant quickly.
4. Start the Gateway
If the daemon was installed, the gateway runs automatically:
Codeopenclaw gateway status
Manual start (foreground):
Codeopenclaw gateway --port 18789 --verbose
Control UI (local browser): http://127.0.0.1:18789/
⚠️ Bun has issues with WhatsApp/Telegram — use Node for these channels.
5. Connect Your Channels
WhatsApp:
Codeopenclaw channels login
- Scan QR code to log in
- First DM will include a pairing code for approval
Telegram / Discord / Mattermost:
- Wizard generates token/config automatically
- First DM requires pairing approval
6. DM Safety (Pairing Approval)
OpenClaw does not process messages from unknown senders until approved:
Check pending pairings:
Codeopenclaw pairing list whatsapp
Approve a code:
Codeopenclaw pairing approve whatsapp <code>
7. Verify End-to-End Functionality
Send a test message:
Codeopenclaw message send --target +15555550123 --message "Hello from OpenClaw"
Check status and health:
Codeopenclaw status --all openclaw health
8. For Developers — Build from Source
Clone and build OpenClaw from GitHub:
Codegit clone https://github.com/openclaw/openclaw.git cd openclaw pnpm install pnpm ui:build pnpm build openclaw onboard --install-daemon
Start the gateway:
Codenode openclaw.mjs gateway --port 18789 --verbose
9. Advanced Usage
- macOS menu bar app + Voice Wake
- iOS/Android nodes (Canvas, camera, voice)
- Remote access (SSH tunnel, Tailscale Serve)
- Always-on setups / VPN environments
Key Benefits of Using OpenClaw
- Local operation: Your data stays under your control
- Multi-channel support: Access via familiar chat apps
- Secure by design: Pairing approval and sandboxed sessions
- Extensible: Add skills or create custom functions via ClawdHub
- Voice & visual features: Natural conversation and screen interaction
OpenClaw combines security, extensibility, and usability in a single, personal AI assistant - and its popularity is growing rapidly in the productivity and developer communities.

Fumi Nozawa
Digital Marketer & Strategist
Following a career with global brands like Paul Smith and Boucheron, Fumi now supports international companies with digital strategy and market expansion. By combining marketing expertise with a deep understanding of technology, he builds solutions that drive tangible brand growth.
Project consultation or other inquiries? Feel free to reach out.
Continue Reading

Why Landing Pages Targeting Japan Fail to Convert for Global SaaS Companies
Why do Japanese landing pages fail to convert for global SaaS companies? Learn the structural differences in information design, trust building, and conversion behavior in Japan.

What is Moltbook?: The AI-Only Social Network Where Agents Self-Organize, Debate, and Govern
What is Moltbook? An AI-only social network where autonomous agents post, debate, and even moderate themselves. Learn how it works and why it matters now.