Skip to content
DLinkClawDLinkClaw Docs

OpenClaw Installation & Model Configuration

This page covers the general OpenClaw installation and how to configure model providers (LLMs) and the default model. Whether you deploy via WSL, a VM, or a physical machine, the configuration is identical.

1. Install OpenClaw

Section titled “1. Install OpenClaw”
Section titled “Option 1: Official installer (recommended)”

macOS / Linux / WSL2:

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

Windows (PowerShell):

Terminal window
iwr -useb https://openclaw.ai/install.ps1 | iex

The script detects your OS, provisions the required Node runtime (Node 24 LTS on Linux, Node 26 on macOS), and installs OpenClaw.

Option 2: npm install

Section titled “Option 2: npm install”

If you manage Node yourself (requires 22.22.3+ / 24.15+ / 25.9+, Node 26 recommended):

Terminal window
npm install -g openclaw@latest --allow-scripts=openclaw
openclaw onboard --install-daemon

Verify

Section titled “Verify”
Terminal window
openclaw --version

2. First-time setup (onboard wizard)

Section titled “2. First-time setup (onboard wizard)”
Terminal window
openclaw onboard

The wizard guides you through:

  1. Model provider: choose your LLM provider and paste your API key
  2. Agent workspace: creates ~/.openclaw/workspace/ as your agent’s home
  3. Chat channels: optionally connect WeChat, Feishu, Discord, DLinkClaw, and more
  4. System service: register the Gateway for auto-start

3. Configure models

Section titled “3. Configure models”

Supported providers

Section titled “Supported providers”

OpenClaw supports major LLM providers including DeepSeek, OpenAI, Anthropic, Google Gemini, Qwen, Doubao (Volcengine), Zhipu, Moonshot, Ollama (local), llama.cpp (local), and more.

Set the default model

Section titled “Set the default model”

List available models:

Terminal window
openclaw models list

Set the default model (provider/model format):

Terminal window
openclaw models set deepseek/deepseek-v4-flash

Example with Volcengine (Doubao):

Terminal window
openclaw models set volcengine-plan/deepseek-v4-flash

Manual config editing

Section titled “Manual config editing”

The config file lives at ~/.openclaw/openclaw.json:

{
  agents: {
    defaults: {
      model: {
        primary: "deepseek/deepseek-v4-flash",
      },
    },
  },
}

Restart the Gateway to apply changes:

Terminal window
openclaw gateway restart

Add more model accounts

Section titled “Add more model accounts”

Run the wizard and choose “Add a provider”, or run openclaw onboard again. Multiple providers can coexist — switch the default anytime with openclaw models set.

4. Gateway service management

Section titled “4. Gateway service management”
Command Description
openclaw status Show Gateway and channel status
openclaw gateway restart Restart the Gateway
openclaw update Update OpenClaw
openclaw uninstall Uninstall OpenClaw

Next steps

Section titled “Next steps”