โ“

OpenClaw Frequently Asked Questions

Clear answers on installation, channels, privacy, skills, runtime requirements, and the browser-based Control UI.

๐Ÿš€

Getting Started

What is OpenClaw? โ–ผ

OpenClaw is an open-source local AI assistant for messaging automation, browser-based Control UI, voice workflows, and skills. Clawdbot remains the legacy brand and search alias, but OpenClaw is the primary product name.

Is OpenClaw free to use? โ–ผ

Yes. OpenClaw is MIT-licensed and free to run. Your ongoing cost depends on the model provider or local infrastructure you choose for the workflows you enable.

Do I need coding skills to install OpenClaw? โ–ผ

You do not need to write code, but you should be comfortable following terminal steps from the install guides. After installation, the official onboarding command is 'openclaw onboard --install-daemon'.

What are the current runtime requirements? โ–ผ

The current docs recommend Node.js 24 and support Node.js 22.14+ for local installs on macOS, Windows, and Linux.

๐Ÿ’ฌ

Channels & Skills

Which channel should I start with first? โ–ผ

Telegram is the fastest starter channel in the official getting-started flow because it only needs a bot token.

How do channel setup commands work? โ–ผ

Use 'openclaw channels add --channel <name>' for token-based or config-driven channels. Use 'openclaw channels login --channel whatsapp' for the interactive WhatsApp Web QR flow.

How do I connect WhatsApp? โ–ผ

Install OpenClaw, run 'openclaw onboard --install-daemon', then use 'openclaw channels login --channel whatsapp' to complete the QR login flow. The WhatsApp setup guide walks through the full sequence.

How do I find and install skills? โ–ผ

Use 'openclaw skills search' to browse the catalog and 'openclaw skills install <slug>' to add a skill. ClawHub is the public registry for discovery and sharing.

๐Ÿ”’

Privacy & Control

Does OpenClaw run locally? โ–ผ

Yes. OpenClaw runs on your own machine, so you decide which channels, providers, and workflows can access your data. Hosted model calls only go to the providers you explicitly configure.

Can I use local or hosted models? โ–ผ

Yes. OpenClaw supports local model setups such as Ollama as well as hosted and OpenAI-compatible providers documented by the project.

How do I open OpenClaw Control UI? โ–ผ

Run 'openclaw dashboard' to open the browser-based Control UI served by the Gateway.

๐Ÿ› ๏ธ

Operations & Troubleshooting

How do I confirm the Gateway is running? โ–ผ

Run 'openclaw gateway status' to check Gateway status, then open 'openclaw dashboard' if you want the browser-based view.

Do I need a separate install for each channel? โ–ผ

No. Install OpenClaw once, then add the channels and skills you need from the same local runtime.

How do I keep installed skills up to date? โ–ผ

Use 'openclaw skills update --all' to refresh installed skills in one pass.

What should I check if a channel stops responding? โ–ผ

First confirm 'openclaw gateway status'. Then reopen 'openclaw dashboard' and review the setup guide for the affected channel to verify credentials, permissions, or interactive login steps are still valid.

Need deeper setup help?

The docs, GitHub project, and community channels cover the current OpenClaw installation and workflow guidance.

Ready to install OpenClaw?

Start with the platform guide that matches your operating system and preferred channel.

Open Installation Guides