Have your own OpenClaw🦞 on AWS / GCP: The Complete Guide

OpenClaw🦞 is your personal AI agent that lives in a cloud computer. It can install software, run code, browse the web, and complete tasks autonomously. In this guide, we'll walk you through getting your own OpenClaw up and running — from zero to a working agent in about an hour.

Overview#

Step	What	Time	Cost
1	Get Your Model API Keys	~30 min	Pay-as-you-go
2	Set Up a Cloud Server	~15 min	~$30–33/mo
3	Install Docker	~5 min	Free
4	Create Your API Key Config	~3 min	Free
5	Pull and Run OpenClaw	~5 min	Free
6	Verify It's Running	~2 min	Free
	Total	~60 min	~$30/mo + model usage

Prerequisites#

Before you begin, make sure you have:

An AWS or GCP account with billing enabled
Basic familiarity with the terminal
A credit card for purchasing model API keys

Step 1: Get Your Model API Keys (~30 min · pay-as-you-go)#

OpenClaw uses your own API keys to call AI models. You'll need to sign up with each provider and grab an API key. You don't need all of them — pick the ones you want to use.

Anthropic (Claude)#

Go to console.anthropic.com
Click Sign Up — you can use Google or email
Go to Settings → Billing → add a credit card and top up credits (minimum $5)
Navigate to API Keys → Create Key
Name your key (e.g. "openclaw") and copy it — starts with sk-ant-...

Model	Input Price	Output Price
Claude Sonnet 4	$3 / 1M tokens	$15 / 1M tokens
Claude Opus 4	$15 / 1M tokens	$75 / 1M tokens
Claude Haiku 4	$0.80 / 1M tokens	$4 / 1M tokens

OpenAI (GPT)#

Go to platform.openai.com
Click Sign Up — you can use Google, Microsoft, or Apple account
Go to Settings → Billing → Add payment details → add a credit card and top up credits (minimum $5)
Navigate to API Keys → Create new secret key
Name your key and copy it — starts with sk-...

Model	Input Price	Output Price
GPT-4o	$2.50 / 1M tokens	$10 / 1M tokens
GPT-4o mini	$0.15 / 1M tokens	$0.60 / 1M tokens
o1	$15 / 1M tokens	$60 / 1M tokens

Google (Gemini)#

Go to aistudio.google.com
Sign in with your Google account
Click Get API Key → Create API key in new project
Copy the key
To use paid models, go to console.cloud.google.com → enable billing and link the project

Model	Input Price	Output Price
Gemini 2.5 Pro	$1.25 / 1M tokens	$10 / 1M tokens
Gemini 2.5 Flash	$0.15 / 1M tokens	$0.60 / 1M tokens

xAI (Grok)#

Go to console.x.ai
Click Sign Up — you can use your X (Twitter) account or email
Go to Billing → add a credit card and top up credits
Navigate to API Keys → Create API Key
Name your key and copy it

Model	Input Price	Output Price
Grok 3	$3 / 1M tokens	$15 / 1M tokens
Grok 3 Mini	$0.30 / 1M tokens	$0.50 / 1M tokens

DeepSeek#

Go to platform.deepseek.com
Click Sign Up — use email or phone number
Go to Top Up → add credits via credit card or other payment methods (minimum ~$2)
Navigate to API Keys → Create new API key
Name your key and copy it

Model	Input Price	Output Price
DeepSeek V3	$0.27 / 1M tokens	$1.10 / 1M tokens
DeepSeek R1	$0.55 / 1M tokens	$2.19 / 1M tokens

Tip: Start with just one or two providers. Claude and DeepSeek are a great combo — Claude for quality, DeepSeek for cost efficiency.

Save your keys somewhere safe. You'll need them in Step 3.

Step 2: Set Up a Cloud Server (~15 min · ~$30–33/mo)#

You need a server to run OpenClaw on. Choose either AWS or GCP.

Option A: AWS (EC2)#

Go to aws.amazon.com and create an account (you'll need a credit card and phone number for verification)
Once logged in, go to EC2 → Launch Instance and configure:

Setting	Recommended Value	Est. Cost
AMI	Ubuntu 24.04 LTS	Free
Instance Type	t3.medium (2 vCPU, 4GB RAM)	~$30/mo
Storage	30GB SSD	~$3/mo
Security Group	Allow ports 22, 80, 443	Free
	Total	~$33/mo

Then SSH into it:

ssh -i your-key.pem ubuntu@your-ec2-ip

Option B: GCP (Compute Engine)#

Go to cloud.google.com and click Get started for free (you'll get $300 free credits for 90 days)
Add a credit card for billing verification
Go to Compute Engine → VM Instances → Create Instance and configure:

Setting	Value	Est. Cost
Machine Type	e2-medium (2 vCPU, 4GB RAM)	~$25/mo
Boot Disk	30GB SSD	~$5/mo
OS	Ubuntu 24.04 LTS	Free
	Total	~$30/mo

gcloud compute instances create openclaw-vm \
  --zone=us-central1-a \
  --machine-type=e2-medium \
  --image-family=ubuntu-2404-lts \
  --image-project=ubuntu-os-cloud \
  --boot-disk-size=30GB \
  --tags=http-server,https-server

Allow HTTP traffic:

gcloud compute firewall-rules create allow-openclaw \
  --allow tcp:80,tcp:443 \
  --target-tags=http-server,https-server

SSH into it:

gcloud compute ssh openclaw-vm --zone=us-central1-a

Step 3: Install Docker (~5 min · free)#

Once you're SSH'd into your server, install Docker:

sudo apt update && sudo apt install -y docker.io docker-compose
sudo usermod -aG docker $USER

Log out and back in for the group change to take effect.

Step 4: Create Your API Key Config (~3 min · free)#

Create a .env file with the API keys you got from Step 1:

.env

ANTHROPIC_API_KEY=sk-ant-xxx
OPENAI_API_KEY=sk-xxx
GOOGLE_API_KEY=xxx
XAI_API_KEY=xxx
DEEPSEEK_API_KEY=xxx

You only need to include the keys for providers you signed up with. For example, if you only got Claude and DeepSeek:

.env

ANTHROPIC_API_KEY=sk-ant-xxx
DEEPSEEK_API_KEY=xxx

Step 5: Pull and Run OpenClaw (~5 min · free)#

docker pull openclaw/openclaw:latest
docker run -d \
  --name openclaw \
  -p 80:8080 \
  --env-file .env \
  --restart unless-stopped \
  openclaw/openclaw:latest

Step 6: Verify It's Running (~2 min · free)#

docker logs openclaw

You should see:

🦞 OpenClaw is running on port 8080
✓ Connected to OpenClaw API
✓ Models loaded: claude, deepseek

Open http://your-server-ip in your browser — your OpenClaw🦞 is ready! Give it a task and watch it go.