Install OpenClaw on Windows: A Complete Step-by-Step Guide

Before You Start

This Duck Editor tutorial will walk you through installing OpenClaw on Windows via WSL, with screenshots for every step.

This tutorial is for installing OpenClaw on Windows 10 (version 2004+) or Windows 11.

💡 Read this first: Why is installing OpenClaw natively on Windows not recommended? — Learn why WSL is the best choice for Windows users.

If you’re on Mac, check out the macOS Installation Guide. If you’d rather not install locally, consider the Cloud Deployment Guide.

What You’ll Need

• Windows 10 (version 2004 or above) or Windows 11
• At least 8GB RAM (16GB recommended)
• At least 5GB of free disk space
• A stable internet connection
• An OpenAI / Google / Anthropic API Key (at least one)

Check your Windows version: Press Win + R, type winver, and press Enter.

---

Choose Your Installation Method

There are two ways to install on Windows:

Method	Difficulty	Best for	Advantage
Method A: Native Windows	⭐ Easy	Beginners	No extra tools needed
Method B: WSL 2	⭐⭐ Medium	Developers	Linux environment, more stable

Beginners should go with Method A. If you have development experience, Method B is recommended.

---

Method A: Native Windows Installation

Step 1: Install Python 3.11+

1. Go to python.org/downloads
2. Download the latest Python 3.11+ installer
3. Run the installer

⚠️ Important! Make sure to check “Add Python to PATH” during installation — many people forget this step!

4. After installation, open Command Prompt (press Win + R, type cmd):

python --version
REM Should display Python 3.11.x or newer

pip --version
REM Should display pip 24.x.x

If python doesn’t work, try python3 or py.

Step 2: Install Git

1. Go to git-scm.com/download/win
2. Download and install (use the default settings)
3. Verify:

git --version
REM Should display git version 2.x.x

Step 3: Install Node.js

1. Go to nodejs.org
2. Download the LTS version
3. Install (use the default settings)
4. Verify:

node --version
REM Should be v18 or above

Step 4: Download OpenClaw

REM Create a project directory
mkdir C:\Projects
cd C:\Projects

REM Download OpenClaw
git clone https://github.com/openclaw/openclaw.git
cd openclaw

Step 5: Create a Virtual Environment

REM Create the virtual environment
python -m venv .venv

REM Activate the virtual environment
.venv\Scripts\activate

REM Once activated, the prompt will show (.venv)
REM (.venv) C:\Projects\openclaw>

⚠️ You need to re-activate every time you open a new Command Prompt:

cd C:\Projects\openclaw
.venv\Scripts\activate

Step 6: Install OpenClaw

REM Upgrade pip
pip install --upgrade pip

REM Install dependencies
pip install -r requirements.txt

⚠️ If you get a C++ compilation error, you need to install Visual Studio Build Tools:

1. Go to visualstudio.microsoft.com/visual-cpp-build-tools/
2. Download and install
3. Select “Desktop development with C++”
4. Re-run pip install -r requirements.txt

Step 7: Set Up Your API Key

REM Copy the config template
copy .env.example .env

REM Edit with Notepad
notepad .env

Fill in at least one API Key in .env:

OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxx
GOOGLE_API_KEY=your-google-api-key
ANTHROPIC_API_KEY=sk-ant-xxxxxxxx

Don’t have an API Key yet? Check out the AI Model API Key Guide — Google AI Studio offers a free tier, perfect for beginners.

Step 8: Launch OpenClaw

python -m openclaw start

If you see the following message, you’re all set:

🐾 OpenClaw is starting...
✅ Server running at http://localhost:3000

Open your browser and go to http://localhost:3000.

If Windows Defender pops up a firewall alert, click “Allow access.”

---

Method B: WSL 2 Installation (Advanced)

WSL (Windows Subsystem for Linux) lets you run Linux inside Windows. The benefit is that many Python packages are more stable on Linux.

📘 Full tutorial: For a deeper understanding of WSL and more usage tips, see the Complete WSL Guide. Below is the quick installation version.

Step 1: Verify BIOS Virtualization is Enabled

WSL 2 requires hardware virtualization support. First, check if your computer has it enabled:

Open Command Prompt as Administrator (Win + R → type cmd → Ctrl + Shift + Enter):

systeminfo

Look at the last few lines and find the “Hyper-V Requirements” section:

Hyper-V Requirements:    A hypervisor has been detected. Features required for Hyper-V will not be displayed.

If it shows “Virtualization Enabled In Firmware: No”, you need to enable it in BIOS:

1. Restart your computer, press DEL or F2 to enter BIOS
2. Find Intel VT-x (Intel processors) or AMD-V / SVM Mode (AMD processors)
3. Set it to Enabled
4. Save and restart

Step 2: (Recommended) Create a Dedicated Windows User

Security tip: Running OpenClaw / AI Agents under a dedicated Windows account limits their system access scope. If you just want to try it out, you can skip this step.

REM Run as administrator
net user openclaw-user YourPassword123 /add

After creating the account, go to Start → User avatar → “Switch user” and log in with the new account.

Step 3: Install WSL 2

Open PowerShell as Administrator (right-click → Run as administrator):

wsl --install

Restart your computer after installation completes.

After restarting, Ubuntu will automatically open and ask you to set up a username and password.

⚠️ If wsl --install doesn’t respond, go back and verify that BIOS virtualization is enabled (Step 1).

Step 4: Install Node.js 22+ (using nvm)

OpenClaw requires Node.js 22 or above. Open Ubuntu (launch it from the Start menu):

# Check existing Node version (if installed)
node -v

# If an old version is installed, remove it first
sudo apt remove nodejs -y
sudo apt autoremove -y
sudo apt update

# Install nvm (Node Version Manager)
curl -fsSL https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.4/install.sh | bash

# Reload config (or exit and re-enter)
exit

Re-enter WSL:

wsl -d ubuntu

# Verify nvm
nvm --version

# Install Node.js 22 (OpenClaw minimum requirement)
nvm install 22
nvm use 22

# Set as default
nvm alias default 22

# Verify
node -v   # Should show v22.x.x
which node  # Should be at ~/.nvm/versions/node/v22.x.x/bin/node

Step 5: Install and Launch OpenClaw

# Option 1: Global npm install (recommended, easiest)
npm install -g openclaw
openclaw --version

# Option 2: Traditional git clone install
sudo apt install python3.11 python3.11-venv python3-pip git -y
mkdir -p ~/Projects && cd ~/Projects
git clone https://github.com/openclaw/openclaw.git
cd openclaw
python3.11 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

Set Up API Key

# npm install method: configure during first launch
openclaw start
# First run will guide you through API Key setup

# git clone method: edit .env
cp .env.example .env
nano .env  # Enter API Key, Ctrl+O to save, Ctrl+X to exit
python -m openclaw start

Don’t have an API Key yet? Check out the AI Model API Key Guide. Or consider the Ollama local option — with ollama launch openclaw you can use free cloud models without any API Key.

WSL’s localhost:3000 is directly accessible from your Windows browser.

🚀 Quick Option: Use Ollama Without an API Key

If you just want to try it out without applying for an API Key, you can install Ollama in WSL:

# Install zstd (required by Ollama)
sudo apt install -y zstd

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Verify Ollama version
ollama --version

# One-click launch OpenClaw + select cloud model
ollama launch openclaw
# Select a cloud model (e.g., kimi-k2.5, minimax-m2.5) → no model download needed, Ollama has a free tier

For more details, see Ollama + OpenClaw Complete Guide.

---

Common Issues

python command not found

Cause: Forgot to check “Add Python to PATH” during installation.

Solution:

1. Re-run the Python installer
2. Select “Modify”
3. Make sure “Add Python to environment variables” is checked

Or manually add to PATH:

setx PATH "%PATH%;C:\Users\YourUsername\AppData\Local\Programs\Python\Python311"

pip install shows Microsoft Visual C++ 14.0 is required

Install Visual Studio Build Tools (see Step 6).

Firewall blocking OpenClaw

Windows Defender may pop up a warning. Click “Allow” — this just lets Python listen on a local port.

localhost won’t open in WSL

Try using WSL’s IP:

hostname -I
# Use the displayed IP instead of localhost

---

Next Steps

Installation complete! Here’s what you can do next:

• 🧩 Learn about Skills: Teach AI repeatable workflows
• 🤖 Build your first Agent
• 🔑 Get a Google API Key to connect more services
• ☁️ Don’t want to manage servers? Try cloud deployment