Build a 24/7 Personal AI Agent with OpenClaw (Clawdbot) on AWS for Free |

Note: The tool mentioned in this article, formerly known by various names including Clawdbot and Cloudbot, has been rebranded to OpenClaw. This article has been updated to reflect the new name and installation instructions. For the latest information, please visit the official website: openclaw.ai.

The tech world is buzzing about OpenClaw (formerly Clawdbot). Many are rushing to purchase new Mac Minis or dedicated hardware just to keep it running 24/7. But there’s a smarter way. You don’t need to spend $600 on new hardware or leave your gaming PC running all night.

The best part? We can host it on the AWS free tier. This creates a cloud-based agent that never sleeps and communicates with you directly through Telegram or WhatsApp. This article will guide you through building your own personal AI agent using OpenClaw.

OpenClaw is a powerful open-source project designed to be the missing link between Large Language Models (LLMs) and your everyday applications.

Understanding the Costs

Before we dive in, let’s be transparent about the costs involved.

OpenClaw itself is free and open-source. The server we’re building is also free, provided you’re eligible for the AWS free tier. The only real cost comes from the agent’s ‘brain’—the LLM it connects to via an API key.

While you pay for what you use per message, we can make this incredibly cost-effective. Instead of expensive credits, we’ll power this agent with the Google Gemini API. Google Cloud currently offers a generous $300 in credits and a 90-day free trial, meaning you can run this powerful AI agent completely free to get started.

Step 1: Prepare the AWS Server

First, log into your AWS console and navigate to the EC2 section.

Launch Instance: Click “Launch instance” to begin configuring the server.
Name Your Server: Give your virtual machine a memorable name, such as my_OpenClaw_vm.
Choose an OS: Select Ubuntu for the operating system.
Select Instance Type: This is a critical step. While a t3.micro instance might seem tempting, its 1GB of memory is insufficient for the installation. It’s advisable to select a virtual machine with at least 4GB of memory. A c7i.large instance, which provides 4GB of memory, is a suitable choice.
Create a Key Pair: You’ll need a key pair to connect to your virtual machine via SSH. Click “Create new key pair,” give it a name, and create it. Your browser will download a .pem file. Keep this file safe.
Network Settings: Ensure that “Allow SSH traffic from” is enabled. You do not need to open any other ports for this setup.
Launch: With everything configured, click “Launch instance” to create your virtual machine.

You have successfully created the server. Now, let’s proceed with the installation.

Step 2: Install OpenClaw

Once the instance is running, it’s time to connect to it and install the software.

Connect via SSH: In your EC2 dashboard, select the instance and click the “Connect” button. The simplest method is to use the “EC2 Instance Connect” option, which opens an SSH session directly in your browser.
Update Packages: Once connected, the first step is to update the system’s packages and install necessary dependencies. Run the following commands one by one:
```
sudo apt update
sudo apt upgrade -y
```
Install OpenClaw: Installing the OpenClaw application is straightforward with a single command.
```
curl -fsSL https://openclaw.ai/install.sh | bash
```
The installation process will take a few minutes to complete.

Step 3: Configuration and Onboarding

With the installation finished, it’s time to configure your new AI agent. The setup wizard will guide you through the process.

Run openclaw in your terminal to start the onboarding.
Select Yes to acknowledge the application’s capabilities.
Choose the Quick start mode.

Connect to an LLM

This is the most important part. You need to select a model and provide an API key. For this demonstration, we will use Google Gemini.

Select Google from the list of providers.
You will be prompted for your Google Gemini API key. To get your key, navigate to aistudio.google.com, go to the “API key” section, and generate a new key.
Paste your API key into the terminal and press Enter.
Next, select a model. To maintain cost-effectiveness, the Flash models are a great choice. While Gemini 1.5 Flash may have compatibility issues, Gemini Pro Flash (or a similar cost-effective version) works well.

Step 4: Connect to a Chat Application

Now, let’s connect the agent to a chat interface. We’ll use Telegram for this guide.

Create a Telegram Bot:
- Open Telegram and search for the BotFather user.
- Start a chat and send the /newbot command.
- Follow the prompts to give your bot a name and a username.
- BotFather will provide you with an API token. Copy this token.
Provide Token: Paste the Telegram bot token into the OpenClaw setup prompt and press Enter.
Skip Skills and Hooks: You can skip configuring skills and hooks for now.
Configure the Bot’s Personality: The setup will ask you a series of questions to define your bot’s personality.
- What is the bot’s name?
- What is its creature?
- What is its conversational tone?
- What is its emoji?
- What is your name?
For example, you could provide the following prompt:

Your name is MyArcadesh and you are my best friend. You should act like that. My name is John, and you can call me John.
Pair with Telegram: The bot will provide a pairing code. Open a chat with your new bot in Telegram and send it this code to complete the pairing process.

Once paired, you can start chatting with your agent directly through Telegram.

User: How is the weather today? I’m in Riyadh and please provide the temperature in Celsius.

MyArcadesh: The weather in Riyadh is currently clear with a temperature of 35°C.

Step 5: Supercharge Your Agent with Integrations

Let’s make the agent even more useful by connecting it to other services, turning it into a second brain.

Connect to Notion

You can instruct your agent to write directly into your Notion workspace.

In your Telegram chat with the bot, send the message: I want to connect to my notion workspace.
The bot will guide you. The process involves creating a new integration in Notion (Settings & Members -> Integrations -> Develop your own integrations).
Generate an API key from your Notion integration and provide it to the bot.
Remember to share the specific Notion page(s) you want the bot to access with the integration you created.

Enhance with Web Search (Brave API)

To give your agent the ability to browse the internet, you can integrate the Brave Search API.

Go to the Brave Search API portal.
Create a free API key.
Provide this API key to your bot when prompted during a similar connection flow.

A Practical Example

Let’s see a practical example of how these integrations work together. Imagine you have an upcoming job interview. You could send your agent the following prompt:

User: Hi MyArcadesh, tomorrow I will have a job interview for a Data Engineer role. Can you please search for frequently asked questions and some guideline documents for me? Please add that information to my Notion workspace. Thanks.

The agent will then use its web search capability to find relevant information and populate your specified Notion page with structured notes, including common interview questions and preparation guidelines. When you check your Notion workspace, you will find a new entry with detailed information ready for your interview preparation.

Final Thoughts

And there you have it. A secure, private, 24/7 AI agent running on AWS. You’ve successfully built a powerful tool tailored to your needs, moving beyond simple subscription services to create something truly your own.