The Ultimate Beginner’s Guide to OpenClaw

- This is the ultimate beginner's guide to OpenClaw,

and by the end you'll have it fully mastered.

We're going from zero to a fully working setup

with deployment, security, costs,

messaging, skills, memory,

automations, model routing, all of it.

And we're doing it without writing a single line of code.

Now, if you've looked at other tutorials

and felt overwhelmed,

or you've tried setting this up already

and hit a wall, stay with me.

I'll walk you through every step,

explain the why behind each one,

and flag the gotchas that trip most people up.

In the next hour,

you'll have your own AI assistant running 24/7,

messaging you on Telegram,

and actually doing things for you,

not just chatting.

Let's get into it.

So what is OpenClaw?

Well, most AI tools are places you go to.

OpenClaw is something that works for you

and can actually come to you.

ChatGPT, Claude, Gemini,

those are all tools you open when you need help,

but OpenClaw is different.

It runs 24/7 on a server, connects to your apps,

and can actually take action without you asking.

There are three pillars to this.

First is brain and memory.

It connects to AI models via API,

remembers everything,

and gets better over time.

Second, it's always on.

It runs 24/7, so it can reach out to you first,

schedule tasks, monitor things,

and send updates.

Then you have tools and actions.

It connects to Telegram, Gmail,

Calendar, Drive, Slack, Discord,

and can actually do things.

Here are some real-world examples

that people have shared online.

"Every morning at seven,

it checks my calendar, scans my email,

and sends me my priorities for the day."

"It saw an interview on my calendar

and automatically created a prep doc,

researched the company and role,

then matched it against my resume."

"My gym only lets you register 24 hours before a class.

I have it watch the schedule

and then register me automatically before classes fill up."

Now you might see it called Clawdbot or Moltbot online.

It's the same project;

it's just been renamed a couple of times.

OpenClaw is the current name,

and that's what we're building with today.

So let's start with where to actually run this thing.

There are three ways to run OpenClaw.

Option one is your personal computer.

This is free and easy to start,

but it stops when your laptop closes.

Your personal files, passwords, browser history,

those are all accessible to the agent.

If something goes wrong,

it's happening on your personal machine.

Not ideal.

Option two is a Mac mini or dedicated spare hardware.

This gives you good isolation

and it's always on if you keep it plugged in.

But if you're buying hardware,

it can cost $500+ upfront.

You need to do port forwarding, deal with power outages,

and then there's internet reliability.

Option three: you can run it on a server or VPS.

This is a separate computer in the cloud.

It starts at just a few bucks per month,

stays online 24/7,

and if OpenClaw breaks or something happens,

it's self-contained.

If things go really wrong,

you can nuke the server and start over.

For most people, this is the right choice.

Now, for this tutorial, we're using Hostinger,

and the reason is simple.

They have a one-click OpenClaw template.

You don't need to use terminal,

you don't need to know Docker.

You just click deploy, and it works for you.

It's really easy.

They also handle

some of the baseline security automatically.

Your gateway gets a randomised port

and pre-configured authentication,

which already puts you ahead

of most of the setups out there.

Use the link on screen

or the first link in the description below.

That'll take you to this page here

and automatically apply an extra 10% discount

on any VPS plan.

By default, the plan is automatically set to KVM 2,

and that'll give you plenty of room to grow:

two CPU cores, 8 gigs of RAM,

100 gigs of disc space.

But if you want to start smaller and cheaper,

you can always switch to the KVM 1 plan and upgrade later.

KVM 1 is enough for a basic setup

where your bot is making API calls

and running a few automations.

However, if you start adding a lot of skills

or running multiple agents at once,

that's when you'd want to upgrade.

Or if you want to run a local model like Ollama,

you'll actually want to grab the KVM 4 plan

so that you have enough RAM to run the model.

For this video, I'm going to go with KVM 1.

Like I said, you can always upgrade later.

When you've got the plan you want,

go ahead and click Deploy.

That'll take you to the cart page here

where, if you look in the order summary,

you can see that the extra 10% discount

has automatically been applied.

Now the first thing you'll need to do

is select the period for your registration.

You can choose between one month,

12 months, and 24 months.

You'll need to choose at least 12 months

to take advantage of our coupon,

but generally you get a better per month price

if you select 24 months.

To get started for the cheapest overall price,

I'm going to choose 12 months for this video.

Next, if you scroll down,

there's a ready to use AI automatically selected,

but I'm going to go ahead and turn this off,

and I recommend you do too.

We can connect our own LLM later,

I'll show you how to do this,

and that'll save you a bit of money.

Next, if you scroll to the bottom,

there's an option for daily auto-backups.

Now, OpenClaw can reconfigure its own environment,

so if something breaks, having daily auto-backups

is like a really powerful undo button

that's worth every penny.

$3 a month is great value for this.

Now, if you wanted to skip this

and get the absolute cheapest startup cost,

leaving this off will get you in for under $70 total.

However, like I said,

I think this is really powerful,

so I'm going to go ahead and turn this on for my video here

just in case anything breaks.

Finally, you'll want to choose a server location.

Generally speaking,

you just want to pick the server location

that has the lowest latency for the fastest speeds.

Once you've got everything configured to your liking,

go ahead and click Continue.

On the next page,

you'll need to register an account.

Go ahead and do that with either Google,

GitHub, or an email address.

And then on the next page

enter your billing information to complete the payment.

After payment,

you'll land on the OpenClaw configuration page,

and I should call out right away

that this might look different

by the time you're following along with this video.

Hostinger pushes changes

to this configuration page pretty regularly.

So fields might be in a different order.

There might be new fields.

Don't worry too much about this,

because the essential fields

that we're going to talk about today should be there,

because they're crucial for setup.

The first thing we're going to talk about

is the OpenClaw gateway token.

That's this first field here right now.

This is your master key to your entire setup.

Anyone with this token has full access to your dashboard

and everything your bot can do.

Click the "i" icon, copy this,

and save it in a password manager.

Don't paste it in a chat,

don't screenshot it,

don't leave it in a random text file.

Treat it like a password,

because it is one.

Next, OpenClaw needs a brain,

and that comes from an AI model through an API key.

I'm going to use Anthropic Claude

because I like how it works,

but you can use OpenAI, Gemini,