OpenClaw Installation Guide: Step-by-Step Setup for Mac and Windows

When I first heard the name OpenClaw, I honestly thought it was some kind of game. Turns out it’s actually a pretty powerful AI automation platform — it lets you hook up all sorts of AI models (DeepSeek, Kimi, Claude, you name it) and connect them directly to chat tools like Telegram and Feishu. Basically, you get to build yourself an AI assistant that’s always on call.

In this article, I’ll walk you through getting OpenClaw up and running. Both Mac and Windows users are covered, so just follow along.

Before We Start

Which OS Should You Use?

OS	Recommendation	Notes
Mac	Highly recommended	Best experience — native access to Calendar, Notes, Screenshots, etc.
Windows	Fully supported	WSL2 recommended for a near-Mac experience
Linux	For tinkerers	If you’re a Linux user, you probably don’t need me to tell you what to do

Bottom line: Use Mac if you have one. Windows works perfectly fine too.

Which Chat Tool?

• International users: Telegram is your best bet — strongest Bot ecosystem, most features
• China-based users: Feishu (Lark) is the way to go — solid API, stable messaging, generous free tier

WeChat Work, DingTalk, and QQ are also supported, just a bit more hassle to configure.

Hardware Requirements

Don’t worry, OpenClaw itself is lightweight — all the AI heavy lifting happens in the cloud:

• CPU: Anything reasonable
• RAM: 8GB minimum (16GB recommended)
• Storage: 10GB free space
• Network: Internet access (you’ll be calling APIs after all)

---

Mac Installation

Installing OpenClaw on Mac is the easiest — basically one command and you’re done.

Step 1: Open Terminal

Press Command + Space, type Terminal, hit Enter. There you go.

iTerm2 or any other third-party terminal works just as well.

Step 2: One-Command Install

Paste this into your terminal and hit Enter:

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

This script will automatically:

• Detect your system environment
• Install Node.js (if you don’t have it)
• Download OpenClaw
• Set up environment variables

The whole thing takes about 2-5 minutes. Go grab a coffee.

Step 3: Verify Installation

openclaw --version

If you see a version number (like 2026.2.9), you’re all set.

Step 4: Initial Configuration

This is the main event. Run the setup wizard:

openclaw onboard

The wizard will walk you through the following:

1. Accept risk notice — Select Yes and move on.

2. Choose startup mode — Pick QuickStart. Keeps things simple.

3. Choose AI model — Here’s the key part:

If you just want to get started quickly, pick qwen! It’ll automatically redirect you to Qwen’s login page — just authorize and you’re done. No API key needed. This is the easiest path for anyone who wants to start using OpenClaw right away.

One caveat though: Qwen’s free tier has a daily question limit. Once you hit it, the bot will let you know. If you find yourself using it a lot, consider setting up your own API key later.

If you’d rather bring your own API key, here are some solid options:

Model	Strength	Best For
DeepSeek	Best value, strong reasoning	Daily chat, code generation
Kimi	2M character context window	Long document analysis
GLM-4	Great Chinese understanding	Chinese content creation
Claude	Top-tier reasoning	Complex tasks, coding
GPT	Well-rounded	General purpose

Heads up: No matter which platform you choose, AI conversations consume tokens — the more you chat, the more you pay. Especially for long texts, file uploads, and analysis tasks, token usage adds up fast. Keep an eye on your usage so you don’t get a surprise bill at the end of the month.

4. Enter API Key (skip this if you chose Qwen) — Go to the model provider’s website, create an API key, and paste it in.

5. Choose chat tool — Feishu, Telegram, or None to skip for now (you can always set it up later).

6. Gateway port — Default 18789 is fine, just hit Enter.

7. Select Skills — Use spacebar to toggle the skill packs you want. Not sure? Just skip.

8. Enable Hooks — Recommended. Used for content guidance, logging, and session recording.

Once the configuration is done, OpenClaw will automatically start the Gateway service.

Step 5: Verify It’s Running

openclaw channels status

If you see Gateway reachable., everything’s working.

Open your browser and go to http://127.0.0.1:18789/chat — you should see the web chat interface.

Everyday Commands

# Start the service
openclaw gateway start

# Enable auto-start on boot (recommended)
openclaw gateway enable

# Stop the service
openclaw gateway stop

# Update OpenClaw
openclaw update

---

Windows Installation

There are two ways to install OpenClaw on Windows. I’ll cover both — pick whichever fits your situation.

Option 1: WSL2 + Ubuntu (Recommended)

This is the officially recommended approach. In plain terms, you run a Linux environment inside Windows, and the experience is basically the same as Mac/Linux.

Step 1: Enable WSL2

Open PowerShell as Administrator (right-click Start menu → Windows PowerShell (Admin)), then run:

dism.exe /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart
dism.exe /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart
wsl --set-default-version 2

After that, restart your computer. Don’t skip this.

Step 2: Install Ubuntu

1. Open Microsoft Store
2. Search for Ubuntu 22.04 LTS (or 24.04 LTS)
3. Click “Get” and install
4. On first launch, set a username and password (this is for the Linux side, pick whatever you want)

Step 3: Update System and Install Dependencies

In the Ubuntu terminal:

sudo apt update && sudo apt upgrade -y
sudo apt install -y curl git wget build-essential

Step 4: Install Node.js

curl -fsSL https://deb.nodesource.com/setup_22.x | sudo -E bash -
sudo apt install -y nodejs

Verify:

node -v   # Should show v22.x.x
npm -v

Step 5: Install OpenClaw

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

Step 6: Verify Installation

openclaw --version

Step 7: Access from Windows

OpenClaw runs inside WSL2, but your Windows browser can access it directly. Start it in the WSL2 terminal:

openclaw gateway run --port 18789

Then open http://localhost:18789 in your Windows browser. That’s it.

For convenience, you can create a start-openclaw.bat file on your desktop:

@echo off
echo Starting OpenClaw Gateway in WSL2...
wsl -d Ubuntu-22.04 -u root service openclaw start
timeout /t 3
start http://localhost:18789

Double-click to launch. Easy.

---

Option 2: Native PowerShell Install

Don’t want to mess with WSL2? Fair enough. You can install directly on Windows.

Step 1: Install Node.js

Go to the Node.js website and download the 22.x LTS Windows installer. Click through the setup.

Make sure to check “Automatically install the necessary tools” during installation.

Step 2: Verify Node.js

Open PowerShell:

node -v
npm -v

Both should output version numbers.

Step 3: Install OpenClaw

Open PowerShell as Administrator, then run:

npm install -g openclaw@latest

Step 4: Fix Permission Issues (If Any)

If you get permission errors, try these:

# Allow script execution
Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

# Change npm directory (avoids permission conflicts)
npm config set prefix "C:\npm"
npm config set cache "C:\npm-cache"

# Add the new directory to PATH
[Environment]::SetEnvironmentVariable("Path", $env:Path + ";C:\npm", "User")

Open a new PowerShell window after making these changes.

Step 5: Verify and Initialize

openclaw --version
openclaw onboard --install-daemon

The configuration wizard is the same as the Mac section — just follow along.

Windows Troubleshooting

sharp module failed to load?

npm cache clean --force
npm install -g openclaw@latest --force

Windows Defender blocking it?

Add these paths to Defender’s exclusion list:

• C:\Users\YourUsername\AppData\Roaming\npm
• C:\Users\YourUsername\.openclaw

---

How to Configure AI Models

OpenClaw doesn’t include a built-in AI model — you need to connect one.

Zero-Config Option: Pick Qwen (Best for Beginners)

As mentioned above, just select qwen in the setup wizard, log in and authorize, and you’re good to go. No API key fiddling required.

The downside is there’s a daily free question limit. If you outgrow it, switch to a self-configured key below.

Self-Configured API Key (Recommended for Long-Term Use)

For a more stable, unlimited experience, get your own API key:

DeepSeek (Best Value)

1. Sign up at DeepSeek Platform
2. Create an API key
3. Select DeepSeek in the OpenClaw wizard and paste your key

Pricing is about 0.001 CNY / 1K tokens. Actual cost depends on your usage.

Kimi (Best for Long Documents)

1. Sign up at Moonshot AI Platform
2. Create an API key
3. Same configuration process

GLM-4 (Best for Chinese)

1. Sign up at Zhipu AI Platform
2. Create an API key
3. Same configuration process

If you already have a Claude or GPT key, those work too — just select the right provider in the wizard.

Reminder: Regardless of which model you use, AI conversations are billed by token. Longer chats, file uploads, and larger contexts all consume more tokens. Check your usage dashboard regularly so you know where you stand.

---

Connecting Chat Tools

Once OpenClaw is installed, connecting a chat tool is what makes it truly useful.

Feishu, Telegram, WeChat Work, QQ — each platform has different setup steps and permission requirements. The good news is the official docs have all platforms covered with detailed steps and permission lists:

Official Channel Configuration Docs: https://docs.openclaw.ai/zh-CN/channels

Click on your platform to find step-by-step setup guides, permission checklists, callback URL configuration, and more.

Here’s a quick overview so you know what to expect:

Telegram (Recommended for International Users)

# Find @BotFather on Telegram, send /newbot to create a bot
# Once you have the Bot Token:

openclaw channels add telegram
openclaw config set channels.telegram.botToken "your-bot-token"
openclaw gateway restart

Feishu / Lark (Recommended for China)

General flow: Create an app on the Feishu Open Platform → Configure event subscriptions → Enable permissions (the official docs list all required ones — don’t miss any) → Bind it in OpenClaw.

# Install Feishu plugin
openclaw plugins install @m1heng-clawd/feishu

# Configure (replace with your own App ID and Secret)
openclaw config set channels.feishu '{"enabled":true,"appId":"cli_xxx","appSecret":"xxx"}' --json

For details on getting your App ID and which permissions to enable, check the official docs under the Feishu section.

Other Platforms

WeChat Work, DingTalk, QQ, and more are all documented in the official channel docs.

---

Summary

Scenario	Recommended Approach
Have a Mac	Local install — one command, done
Windows user	WSL2 + Ubuntu — near-native experience
Don’t want WSL	Native PowerShell install — works fine
Need 24/7 uptime	Consider cloud deployment

All in all, setting up OpenClaw isn’t complicated. The core flow is just three steps: Install → Run the wizard → Configure AI model. Everything else is optional based on your needs.

If you run into issues, check the OpenClaw docs or ask in the community.

Happy building.