Install OpenClaw in Minutes: One‑Click Script, npm, and Cloud Options

What is OpenClaw?

OpenClaw is an open‑source AI assistant platform that runs large language models (LLMs) locally or in the cloud. This guide explains how to install OpenClaw on a personal computer.

Prerequisites

Operating system: Windows 10/11, macOS 10.15+, or Linux Ubuntu 18.04+

Memory: ≥ 8 GB (16 GB recommended)

Disk space: ≥ 2 GB free

Network: ability to reach GitHub and Chinese AI model providers (e.g., Zhipu AI, Qwen, MiniMax)

Open a terminal

Windows : press Win+R, type cmd and press Enter.

PowerShell (recommended on Windows) : search “PowerShell” in the start menu and open it.

macOS : press Cmd+Space, type “Terminal” and press Enter.

Linux : press Ctrl+Alt+T (Ubuntu/Debian) or open “Terminal” from the applications menu.

Tip: After the terminal opens you will see a blinking cursor, which means the shell is waiting for your command.

Check / install Node.js (required for the npm method)

Run node -v. If a version string such as v18.17.0 appears, Node.js is already installed. Otherwise install it as follows:

Windows

Visit the official Node.js website (https://nodejs.org).

Download the LTS installer.

Run the installer with default settings.

Re‑open the terminal and verify with node -v.

macOS

Download the .pkg installer from the Node.js website.

Run the installer.

Alternatively, install via Homebrew: brew install node.

Linux

sudo apt update && sudo apt install -y nodejs npm   # Ubuntu/Debian

sudo yum install -y nodejs npm   # CentOS/RHEL

Why Node.js? The OpenClaw npm package depends on Node.js to execute JavaScript scripts.

One‑click script installation (recommended for beginners)

Copy and paste the following command in the terminal:

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

The script performs four actions:

Detects the system environment.

Downloads the latest OpenClaw release.

Installs required dependencies.

Configures the runtime environment.

When the script finishes you will be prompted to select a large‑model provider and paste your API key (e.g., zhipuai-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx).

Complete the three‑step startup process:

# 1. Run the configuration wizard (first‑time only)

openclaw onboard

During the wizard you will confirm risk warnings, choose a running mode (web, bot, etc.), select a model provider, configure message channels (QQ, Feishu, etc.) and enable desired skills (weather, calendar, news, …).

# 2. Start the gateway service

openclaw gateway start

The gateway is the “switchboard” that receives messages from all channels and forwards them to the appropriate processors.

# 3. (Optional) Open the web dashboard

openclaw dashboard

The dashboard runs at http://127.0.0.1:18789 and shows service status, installed skills, logs and configuration panels.

Verify the installation

Web dashboard : run openclaw dashboard and check that the page loads with status “running”.

Test command in a chat channel : send /help in QQ or Feishu; the bot should reply with a help message.

Simple skill test : send 现在几点了？ (What time is it now?) and verify that OpenClaw returns the current time.

Service status : run openclaw status. Expected output:

[INFO] OpenClaw service status: running
[INFO] Web service: http://127.0.0.1:18789
[INFO] QQ bot: connected
[INFO] Feishu bot: connected

Configure message channels

QQ bot

Install the official QQBot plugin: openclaw plugins install @sliverp/qqbot@latest Add the QQ bot token (AppID:AppSecret):

openclaw channels add --channel qqbot --token "YOUR_APPID:YOUR_APPSECRET"

Restart the gateway:

openclaw gateway restart

Token is obtained from the QQ Bot platform (https://q.qq.com/qqbot/openclaw/login.html).

Feishu bot

Install the official Feishu plugin: npx -y @larksuite/openclaw-lark-tools install During the interactive setup choose to create a new bot (QR‑code scan) or bind an existing one.

After setup, the bot is ready to receive messages in Feishu.

Other installation methods (brief)

npm installation (for developers)

# Ensure Node.js is installed

node -v

# Install OpenClaw globally

npm install -g openclaw@latest

# Run the configuration wizard

openclaw onboard

# Start the gateway

openclaw gateway start

# (Optional) Open the dashboard

openclaw dashboard

Advantages: easy updates, full control. Disadvantages: requires command‑line knowledge and manual environment setup.

macOS client

Download the latest .dmg from the OpenClaw GitHub releases page (https://github.com/openclaw/openclaw/releases/latest).

Drag OpenClaw into the Applications folder.

Launch from Launchpad.

Only macOS is supported.

Tencent Cloud one‑click deployment

Open the Tencent Cloud OpenClaw deployment page (https://cloud.tencent.com/act/pro/openclaw).

Log in, select a server (e.g., 2 CPU + 4 GB RAM), and click “One‑Click Deploy”.

Wait 5–10 minutes for the instance to become ready.

Provides a 24/7 service without local resources.

Alibaba Cloud quick deployment

Open the Alibaba Cloud OpenClaw page (https://www.aliyun.com/solution/tech-solution/clawdbot).

Log in and follow the wizard to provision resources.

Wait for deployment to finish.

Enterprise‑grade support and higher reliability.

Classic use case – daily briefing

After installation you can ask OpenClaw to send a morning report:

每天早上 8 点，把今天的天气、3 条重要新闻、今日日程发给我。

The engine will:

Fetch the weather (weather skill).

Collect three news items (news skill).

Read today’s calendar events (calendar skill).

Format the information and send it to the configured channel.

Common questions & troubleshooting

Q1: One‑click script fails

Network issue : the script downloads from GitHub; retry or use a VPN.

Permission problem : run the command with sudo (e.g., sudo curl -fsSL https://openclaw.ai/install.sh | bash).

Incompatible system : switch to another method (npm or client).

Q2: npm install cannot find the command

Ensure Node.js is installed ( node -v).

Install OpenClaw with the exact package name: npm install -g openclaw@latest Verify installation: openclaw -v Start with the three‑step flow ( openclaw onboard, openclaw gateway start, optional openclaw dashboard).

Q3: macOS client cannot be opened

macOS may block unsigned apps. Allow the app in “System Preferences → Security & Privacy”.

If the .dmg is missing, download it from the GitHub releases page.

Q4: How to update OpenClaw

One‑click script : re‑run the script.

npm : npm update -g openclaw.

macOS client : check for updates inside the app or download the newest .dmg.

Q5: “openclaw start” not found

OpenClaw does not have a single start command. Use the three‑step flow: openclaw onboard – run the interactive configuration wizard (first‑time only). openclaw gateway start – launch the core services (run after each reboot). openclaw dashboard – open the web UI (optional).

To enable automatic start on boot, configure a system service (e.g., systemd) or use the built‑in startup options.

Diagnostic tool

OpenClaw provides a built‑in doctor command to detect and fix common problems:

openclaw doctor          # Run diagnostics

openclaw doctor --fix   # Automatically repair detected issues

The tool checks environment, configuration, service status, skill health and dependencies, and can automatically restart services or reinstall missing packages.

Advanced configuration (optional)

The main configuration file is located at ~/.openclaw/openclaw.json. A typical content block looks like:

{
  "models": {
    "providers": {
      "zhipuai": {
        "baseUrl": "https://open.bigmodel.cn/api/paas/v4",
        "apiKey": "YOUR_ZHIPUAI_API_KEY",
        "models": ["glm-4-flash"]
      },
      "qwen": {
        "baseUrl": "https://dashscope.aliyuncs.com/compatible-mode/v1",
        "apiKey": "YOUR_QWEN_API_KEY",
        "models": ["qwen-turbo"]
      }
    },
    "default": "zhipuai"
  },
  "channels": {
    "qqbot": {
      "token": "YOUR_QQ_APPID:YOUR_APPSECRET",
      "adapter": "websocket"
    },
    "feishu": {
      "app_id": "YOUR_FEISHU_APP_ID",
      "app_secret": "YOUR_FEISHU_APP_SECRET"
    }
  },
  "skills": {
    "enabled": ["weather", "calendar", "news", "browser"]
  },
  "gateway": {
    "port": 18789,
    "host": "127.0.0.1"
  },
  "memory": {
    "conversation_history": 50
  }
}

After editing, restart the gateway: openclaw gateway restart It is recommended to back up the file before modification:

cp ~/.openclaw/openclaw.json ~/.openclaw/openclaw.json.bak
openclaw validate-config

For most users the default configuration works out of the box.

Summary of the installation workflow

Choose an installation method (one‑click script is the simplest for beginners).

Obtain a large‑model API key from a provider such as Zhipu AI, Qwen, or MiniMax.

Run the three‑step startup process: openclaw onboard – interactive wizard. openclaw gateway start – launch services. openclaw dashboard – optional web UI.

Verify the installation via the dashboard, test commands, or openclaw status.

References

Node.js official site: https://nodejs.org

OpenClaw installation docs: https://docs.openclaw.ai/installation

One‑click script: https://openclaw.ai/install.sh

GitHub releases: https://github.com/openclaw/openclaw/releases/latest

QQ Bot platform (token acquisition): https://q.qq.com/qqbot/openclaw/login.html

Feishu plugin guide: https://bytedance.larkoffice.com/docx/MFK7dDFLFoVlOGxWCv5cTXKmnMh

Tencent Cloud deployment page: https://cloud.tencent.com/act/pro/openclaw

Alibaba Cloud deployment page: https://www.aliyun.com/solution/tech-solution/clawdbot

Zhipu AI API docs: https://open.bigmodel.cn/dev/api