How to Run Your First AI Agent With OpenClaw

How to set up OpenClaw for beginners with a simple guide to install OpenClaw, test the dashboard, connect one chat app, and run your first safe AI workflow.

Run OpenClaw Now

What Is OpenClaw and Why Use It?

OpenClaw is an AI agent framework that helps you build an AI assistant that can do real tasks, not just answer questions.

A normal chatbot usually replies inside one app. OpenClaw can connect with different AI models, tools, files, browsers, and chat apps like Telegram, WhatsApp, Discord, and Slack. This allows your assistant to work across the tools you already use.

You can use OpenClaw for tasks like:

Use Case	What OpenClaw Can Do
Reminders	Send task reminders or follow-up alerts
Research	Search, collect, and summarize information
File work	Read, organize, rename, or summarize files
Calendar tasks	Help manage meetings, schedules, and follow-ups
Workflow automation	Run repeated tasks based on your rules
Chat-based assistant	Let you control workflows from apps like Telegram or WhatsApp

What You Need Before Installing OpenClaw

Before installing OpenClaw, make sure you have these things ready:

Supported system:

Use macOS, Linux, or Windows. For Windows, WSL2 is recommended for a better setup experience.

Node.js installed:

OpenClaw recommends Node 24. It also supports Node 22.16+. Check your version with: node --version

AI model API key:

You need an API key from a model provider like Anthropic, OpenAI, Google, or another supported provider. OpenClaw onboarding will ask for it.

Terminal access:

You need Terminal, PowerShell, or WSL2 terminal to run the install and setup commands.

Stable internet connection:

Required to download OpenClaw, connect model providers, and use online tools.

One safe workspace folder:

Use a dedicated folder for OpenClaw files and workflows. Do not give full computer access at the beginning.

Hardware Recommendations for OpenClaw

Use Case	Recommended Hardware	Best For
Basic testing	2 CPU cores, 2 to 4 GB RAM, 10 GB storage	Trying OpenClaw locally
Beginner personal use	2 CPU cores, 4 GB RAM, 20 GB SSD	Reminders, research, notes, simple workflows
Regular workflows	2 to 4 CPU cores, 8 GB RAM, 20 GB+ SSD	Daily assistant, chat channels, file tasks
Docker setup	2 to 4 CPU cores, 8 GB RAM, 20 GB+ SSD	Safer isolated setup
Browser automation	4 CPU cores, 8 GB+ RAM, 30 GB+ SSD	Web research, browser tasks, screenshots
24/7 VPS setup	2 to 4 CPU cores, 4 to 8 GB RAM, 20 GB+ SSD	Always-on OpenClaw assistant

How to Set Up OpenClaw for Beginners

Follow these steps to install OpenClaw, run onboarding, open the dashboard, and test your first message.

Check What You Need

Before setup, make sure you have:

Node.js installed
API key from a model provider like Anthropic, OpenAI, Google, or another provider

Check your Node version:

node --version

Install OpenClaw

macOS / Linux

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

Windows PowerShell

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

Run OpenClaw Onboarding

After installation, run:

openclaw onboard --install-daemon

The onboarding wizard helps you:

choose a model provider
add your API key
configure the Gateway
set up basic authentication
prepare OpenClaw to run in the background

Verify the Gateway Is Running

Run:

openclaw gateway status

If everything is working, you should see the Gateway listening on port 18789.

The Gateway is the service that lets OpenClaw receive messages, connect tools, and run your assistant.

Open the Dashboard

Run:

openclaw dashboard

This opens the OpenClaw Control UI in your browser.

Send Your First Message

In the Control UI, type a simple message like:

Hello, what can you do?

If OpenClaw replies, the setup is complete.

Connect a Chat Channel

After the dashboard works, connect one chat app.

For beginners, Telegram is the easiest option because it only needs a bot token.

You can also connect Telegram, WhatsApp, Discord, Slack, Microsoft Teams, and other supported channels.

What to Do After Setup

After OpenClaw is running:

connect one channel
set pairing and safety rules
configure the Gateway if needed
browse available tools
enable only the tools you need
create one simple workflow first

Start with one workflow like reminders, research summaries, or daily planning.

Easiest Way to Run OpenClaw for Beginners

If you want to avoid the technical setup, you can use a managed OpenClaw hosting platform like Ampere.sh instead of installing everything yourself.

Go to ampere.sh and create an account.
Click Deploy Agent to launch your OpenClaw agent.
Connect your preferred AI model, or use the included credits available with your Ampere plan.
Connect one chat app first, such as Telegram, WhatsApp, or Discord.
Send a test message to make sure your agent is working.
Start your first workflow, such as reminders, email summaries, calendar updates, research, or task planning.

Deploy OpenClaw Now

Common OpenClaw Setup Problems and Fixes With Commands

Problem	Solution	Command
Dashboard not opening	Check if Gateway is running	`openclaw gateway status`
Dashboard still not opening	Open dashboard manually	`openclaw dashboard`
Gateway not running	Restart the Gateway	`openclaw gateway restart`
Port 18789 already used	Check which process is using the port	`lsof -i :18789`
Port issue on Linux	Stop the process using the port	`kill -9 <PID>`
Node.js version error	Check Node version	`node --version`
Telegram/chat pairing issue	Approve pairing code	`openclaw pairing approve telegram <PAIRING_CODE>`
Permission error	Fix folder ownership	`sudo chown -R $USER:$USER ~/.openclaw`
Docker permission error	Fix Docker data folder ownership	`sudo chown -R 1000:1000 ~/.openclaw`
Need health check	Run OpenClaw doctor	`openclaw doctor`
Need auto-fix	Run doctor fix	`openclaw doctor --fix`
Remote dashboard not opening	Use SSH tunnel	`ssh -N -L 18789:127.0.0.1:18789 user@your-server-ip`

Frequently Asked Questions

Is OpenClaw beginner-friendly?

Yes. Beginners can use OpenClaw by running the installer, completing onboarding, connecting one model, and testing the dashboard first.

What do I need before installing OpenClaw?

You need Node.js, an AI model API key, terminal access, internet connection, and a supported system like macOS, Linux, or Windows with WSL2.

Do I need an API key for OpenClaw?

Yes. You need an API key from a model provider like Anthropic, OpenAI, Google, or another supported provider.

What is the OpenClaw Gateway?

The Gateway is the service that runs OpenClaw and lets it receive messages, connect tools, and control workflows.

What is the easiest chat app to connect with OpenClaw?

Telegram is usually the easiest first option because it only needs a bot token.

Can OpenClaw run 24/7?

Yes. For 24/7 use, run it on a VPS, Mac Mini, Docker server, always-on computer, or managed hosting platform.

Also Read

Installation

How to Set Up OpenClaw on Ubuntu and Connect Your AI Agent

Installation

How to Install OpenClaw on a VPS (Step-by-Step Guide)

Integration

Openclaw On Docker

Written by

Michael Park

Senior Technical Writer & DevRel

Michael creates comprehensive installation and setup guides for developers and system administrators. With experience across Linux, macOS, Windows, and embedded systems, he has written over 200 technical tutorials used by millions of developers. He focuses on clear, step-by-step instructions that work the first time, covering everything from Raspberry Pi to enterprise servers.

Easiest Way to Run OpenClaw Without Local Setup

If you do not want to install OpenClaw locally or manage servers, Docker, ports, updates, and logs, Ampere.sh gives you an easier way to run OpenClaw.

Run OpenClaw Now