OpenClaw Install Requirements Checklist

Use this checklist to verify you have everything needed before installing OpenClaw.

Required

Essential Requirements

Optional

Platform Integrations

1 Core OpenClaw Install Requirements

📦

Node.js 22 or Newer

Required

Why it's needed: OpenClaw is built on Node.js and acts as a local agent runtime and message router. Node.js 22+ is required for compatibility with the latest OpenClaw features.

Check Your Version

bash
node --version

Expected output: v22.x.x or higher

Installation Links

📥

npm Package Manager

Required

Why it's needed: npm is used to install OpenClaw globally and manage its dependencies. npm is included automatically when you install Node.js.

Verify npm Installation

bash
npm --version

Expected output: 10.x.x or higher

Install OpenClaw with npm

bash
npm install -g openclaw@latest

2 OpenClaw Platform Requirements

OpenClaw supports macOS, Linux, and Windows. Check the specific requirements for your operating system.

🍎

macOS Requirements

  • OS Version: macOS 11 (Big Sur) or newer
  • Architecture: Intel x86_64 or Apple Silicon (M1/M2/M3/M4)
  • RAM: 4GB minimum, 8GB recommended
  • Disk Space: 1GB for OpenClaw + dependencies
  • Optional: Homebrew package manager
  • Optional: Xcode Command Line Tools
macOS Install Guide →
🐧

Linux Requirements

  • Distributions: Ubuntu 20.04+, Debian 11+, Fedora 35+, Arch
  • Architecture: x86_64 (amd64) or ARM64
  • RAM: 4GB minimum, 8GB recommended
  • Disk Space: 1GB for OpenClaw + dependencies
  • Optional: systemd for service management
  • Optional: Build essentials (gcc, make, python3)
Linux Install Guide →
🪟

Windows Requirements

  • OS Version: Windows 10 (1903+) or Windows 11
  • Architecture: 64-bit (x64)
  • RAM: 4GB minimum, 8GB recommended
  • Disk Space: 1GB for OpenClaw + dependencies
  • Optional: Windows Build Tools
  • Optional: PowerShell 7+
Windows Install Guide →

3 OpenClaw AI Model Requirements

OpenClaw is model-agnostic and supports multiple AI models. You need at least one of the following options to use OpenClaw effectively.

OpenAI GPT

Cloud-Based

GPT models provide state-of-the-art performance with advanced coding capabilities.

Supported Models (2026):

  • GPT-5 - Latest flagship model
  • GPT-5.3-Codex - Most capable agentic coding model (released Feb 5, 2026)

How to Get API Key:

  1. Sign up at platform.openai.com
  2. Create an API key
  3. Add billing information
  4. Use the key during OpenClaw setup

Pricing: Pay-per-use, varies by model

Free Option

Local Models (Ollama)

Local / Free

Run AI models locally on your computer with Ollama integration. No API costs, complete privacy.

Requirements:

  • Models with ≥64,000 token context
  • 8GB+ RAM for smaller models
  • 16GB+ RAM for larger models
  • GPU recommended for faster inference

Recommended Local Models:

  • glm-5 - Zhipu AI's latest open model
  • qwen3-coder - Qwen series coding specialist
  • gpt-oss:20b / 120b - Open-source GPT variants
  • kimi-k2.5 - Local deployment of Moonshot's model

Setup:

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

# Launch OpenClaw with Ollama
ollama launch openclaw

Cost: Free (no API fees)

Chinese AI Models

Cloud-Based

Advanced Chinese AI models with strong multilingual and coding capabilities.

Supported Models (2026):

  • Kimi K2.5 - Moonshot AI, 1T params, multimodal
  • GLM 5 - Zhipu AI, improved coding abilities
  • MiniMax 2.5 - Industry-leading multi-language coding
  • DeepSeek series - Custom configurations

Pricing: Competitive rates, varies by provider

OpenRouter Auto

Intelligent Routing

Built-in OpenRouter integration that automatically selects the most cost-effective model based on task complexity.

Features:

  • Automatic model selection
  • Cost optimization
  • Simple tasks → cheaper models
  • Complex tasks → capable models

Configuration:

Use model ID: openrouter/openrouter/auto

More info: OpenRouter Docs

4 Optional OpenClaw Requirements

These are optional but enhance OpenClaw's capabilities. You can add them during or after installation.

💬

Messaging Platform Tokens

Connect OpenClaw to messaging platforms to interact with your AI assistant from anywhere.

Supported Platforms (50+ integrations):

  • Discord - Bot token from Discord Developer Portal
  • Telegram - Bot token from @BotFather
  • WhatsApp - Session via QR code scan
  • Slack - App token from Slack API
  • Microsoft Teams - Bot registration
  • Signal - Phone number registration
  • iMessage - via BlueBubbles integration
  • Google Chat, Matrix, Zalo - Platform-specific setup

Note: You can also use the web interface at http://127.0.0.1:18789/ without any messaging platform.

🔌

Skills & Plugins

Extend OpenClaw with community-built skills from the ClawHub marketplace.

OpenClaw Ecosystem:

  • 5,705 skills on ClawHub marketplace (Feb 7, 2026)
  • 53 official bundled skills included
  • 25 core tools for functionality
  • 700+ total skills in ecosystem

Skill Resources:

Security: All ClawHub skills are scanned with VirusTotal (v2026.2.6+)

🧠

Memory & Persistence

Enable persistent memory for OpenClaw to remember past interactions and adapt to your habits.

Memory Features:

  • Voyage AI integration - Advanced memory support
  • Persistent memory - Recalls interactions over weeks
  • User adaptation - Learns your preferences
  • Context retention - Maintains conversation history

Configure memory during openclaw onboard or in settings later.

🎭

SOUL.md Personality System

Customize your AI agent's personality with the revolutionary SOUL.md configuration system.

Personality Customization:

  • Define personality - Formal, conversational, or sassy
  • Set values - Agent's beliefs and principles
  • Communication style - Tone, brevity, humor
  • Behavior rules - How agent responds to situations

Community Resources:

⚙️ OpenClaw Hardware Recommendations

Minimum

Basic Setup

For running OpenClaw with cloud-based AI models (Claude, GPT)

  • CPU: Dual-core 2.0GHz+
  • RAM: 4GB
  • Storage: 5GB free space
  • Network: Broadband internet

Suitable for web interface and messaging platform integration

Advanced

Local AI Setup

For running local AI models with Ollama (no API costs)

  • CPU: 8-core 3.0GHz+ or Apple Silicon
  • RAM: 16GB+ (32GB for large models)
  • Storage: 50GB+ free space (NVMe SSD)
  • GPU: NVIDIA RTX 3060+ or AMD equivalent (recommended)

Required for local model inference with good performance

Ready to Install OpenClaw?

Once you've verified all requirements, proceed with the installation guide for your platform.