How to Install OpenClaw – [Beginner Step-by-Step]

Imagine messaging your computer and having it actually do things read files, send emails, or run tasks just like magic. Sounds futuristic, right? Well, it’s here, and it’s called OpenClaw.

I’ve been testing this AI agent for days, and in this guide, I’ll walk you through exactly how to install OpenClaw on your PC. Whether you’re a tech enthusiast or just curious about AI automation, I’ve got you covered with both paid and completely free methods.

Let’s dive in.

What Exactly Is OpenClaw?

OpenClaw is an AI agent that lives on your computer and takes commands via messaging apps like WhatsApp or Telegram. Think of it as your personal digital assistant that can:

• Read and write files on your PC
• Send emails on your behalf
• Perform system tasks
• Respond to your messages in real-time

You literally message a WhatsApp bot, and it executes tasks on your computer. No complicated dashboards. No technical jargon. Just you, your messages, and your PC doing the heavy lifting.

Method 1: Install OpenClaw Using API (Quick Start)

This method uses OpenAI’s API and takes about 5 minutes. It’s perfect if you want to test OpenClaw immediately.

Step 1: Install Node.js

OpenClaw requires Node.js to run. Head over to nodejs.org and download the installer for your operating system (Windows, macOS, or Linux). The installation is straightforward just click through the prompts.

Pro tip: If you’re unsure whether you have Node installed, open your terminal and type node --version. If you see a version number, you’re good to go.

Step 2: Install OpenClaw

Visit openclaw.ai and scroll down until you see the installation command. Copy it, open your terminal, and paste:

curl -fsSL https://openclaw.ai/install | bash

Wait a few seconds, and you’ll see “OpenClaw successfully installed.” Easy, right?

Step 3: Get Your OpenAI API Key

Since OpenClaw needs an AI model to process your requests, you’ll need an API key:

1. Go to platform.openai.com/api-keys
2. Set up billing (don’t worry—I used GPT-4.1 Nano for days and spent almost nothing)
3. Click “Create new secret key”
4. Name it something like “OpenClaw” and copy the key

Step 4: Configure OpenClaw

Back in your terminal, OpenClaw will walk you through setup:

• Choose “Quick Start” when prompted
• Select “OpenAI” as your provider
• Paste your API key
• Pick a model—I recommend GPT-4.1 Nano (it’s incredibly cheap and works great for testing)

Step 5: Connect WhatsApp

Here’s where the magic happens. OpenClaw will ask how you want to access your bot. Select “WhatsApp” and choose “Connect via QR code.”

Grab your phone, open WhatsApp, go to Settings > Linked Devices, and scan the QR code displayed in your terminal. Within seconds, you’ll see “Web session ready.”

Enter your phone number when prompted, and boom—your OpenClaw bot is alive!

Step 6: Test It Out

Open WhatsApp, find your own number (you message yourself to interact with the bot), and type “Hi.”

You should get a response like: “Hello! I’m your OpenClaw assistant, ready to help.”

Try asking a simple question. The bot will think for a moment and respond using the AI model you configured.

Method 2: Run OpenClaw Completely FREE with Ollama

Don’t want to pay for API calls? No problem. You can run OpenClaw locally using Ollama with open-source models. Here’s how:

Step 1: Install Ollama

Visit ollama.com and download the installer for your operating system. It’s free and open-source.

Step 2: Choose and Download a Model

OpenClaw works well with various local models. For beginners, I recommend starting with Llama 3.2 (1B parameters)—it’s lightweight and runs on most computers.

In your terminal, run:

ollama run llama3.2:1b

This downloads the model. If you have more powerful hardware and want better results, consider GLM-4-9B-Chat (about 5GB) or Qwen2.5-7B.

Step 3: Launch OpenClaw with Ollama

Once your model is downloaded, run:

ollama run openclaw

OpenClaw will detect your local models and ask which one to use. Select your downloaded model, and you’re running entirely free—no API keys, no billing, no surprises.

OpenClaw Web UI vs. Terminal

OpenClaw gives you two ways to interact:

Terminal Mode: Type openclaw in your terminal for a command-line interface. Great for quick tasks and testing.

Web UI: During setup, choose “Open the web UI” to access a browser interface where you can monitor connected channels (WhatsApp, Telegram, etc.) and see your bot in action.

I personally use both—the web UI for monitoring and terminal for quick commands.

What Can You Actually Do With OpenClaw?

Once installed, OpenClaw becomes more powerful through “skills” additional capabilities you can enable:

• Apple Reminders – Create reminders via WhatsApp
• Block Watcher – Monitor system blocks
• Email Integration – Read and send emails
• Google Places – Location-based services
• Image Generation – Create images using OpenAI or local models
• Notion Integration – Update your notes
• ElevenLabs – Generate audio

During setup, you can skip these and add them later as you explore.

Troubleshooting Common Issues

“Command not found” after installation?
Restart your terminal or run source ~/.bashrc (or ~/.zshrc for macOS).

Node.js errors?
Make sure you have Node version 18 or higher installed.

WhatsApp not connecting?
Try scanning the QR code again. Ensure your phone has a stable internet connection.

Model running slow?
Local models require decent hardware. Stick with smaller models like Llama 3.2 1B if you’re on an older machine.

Final Thoughts

OpenClaw bridges the gap between messaging apps and your computer in a way that feels genuinely futuristic. Whether you’re automating tasks, experimenting with AI, or just want to impress your friends by controlling your PC via WhatsApp, this tool delivers.

I’ve been running it with GPT-4.1 Nano for days with minimal cost, and the free Ollama method means anyone can try it without spending a dime.

Have questions? Drop them in the comments. And if you build something cool with OpenClaw, I’d love to hear about it.

Happy automating!