Cloud Development Intermediate

Develop with Claude Code on cloud platforms — from major cloud providers like AWS, GCP, and Azure to cloud IDEs like GitHub Codespaces and Gitpod. Learn setup, configuration, and cost optimization for each platform.

AWS EC2 + Claude Code

Terminal

# Launch an EC2 instance (Amazon Linux 2023 or Ubuntu)
# Install Node.js and Claude Code
$ ssh -i my-key.pem ec2-user@ec2-xx-xx-xx-xx.compute.amazonaws.com

# Install Node.js via nvm
$ curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.7/install.sh | bash
$ source ~/.bashrc
$ nvm install 20

# Install Claude Code
$ npm install -g @anthropic-ai/claude-code

# Use AWS Secrets Manager for API key
$ export ANTHROPIC_API_KEY=$(aws secretsmanager get-secret-value \
    --secret-id claude-api-key \
    --query SecretString --output text)

# Start developing
$ cd /home/ec2-user/my-project
$ claude

AWS Cloud9: AWS Cloud9 provides a browser-based IDE with a built-in terminal. Install Claude Code in the Cloud9 terminal for a fully cloud-based development experience without SSH.

Google Cloud Shell

Google Cloud Shell provides a free, ephemeral VM with 5GB of persistent storage:

Terminal

# Google Cloud Shell already has Node.js
# Open Cloud Shell from console.cloud.google.com

# Install Claude Code
$ npm install -g @anthropic-ai/claude-code

# Store API key in Secret Manager
$ export ANTHROPIC_API_KEY=$(gcloud secrets versions access latest \
    --secret=claude-api-key)

# Clone your project and start coding
$ git clone https://github.com/your-org/project.git
$ cd project
$ claude -p "Set up the development environment and run the app"

Azure VM

Terminal

# SSH into your Azure VM
$ ssh azureuser@my-vm.eastus.cloudapp.azure.com

# Install Node.js
$ curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
$ sudo apt-get install -y nodejs

# Install Claude Code
$ sudo npm install -g @anthropic-ai/claude-code

# Use Azure Key Vault for API key
$ export ANTHROPIC_API_KEY=$(az keyvault secret show \
    --vault-name my-vault \
    --name claude-api-key \
    --query value -o tsv)

$ claude

GitHub Codespaces

GitHub Codespaces provides a full development environment in the cloud, accessible from your browser:

JSON (.devcontainer/devcontainer.json)

{
  "name": "Claude Code Development",
  "image": "mcr.microsoft.com/devcontainers/javascript-node:20",
  "postCreateCommand": "npm install -g @anthropic-ai/claude-code && npm install",
  "secrets": {
    "ANTHROPIC_API_KEY": {
      "description": "Your Anthropic API key for Claude Code"
    }
  },
  "customizations": {
    "codespaces": {
      "openFiles": ["CLAUDE.md"]
    }
  }
}

Codespaces Secrets: Add your ANTHROPIC_API_KEY as a Codespace secret in your GitHub repository settings (Settings → Secrets and Variables → Codespaces). It will be automatically available as an environment variable.

Gitpod

YAML (.gitpod.yml)

image:
  file: .gitpod.Dockerfile

tasks:
  - name: Setup
    init: |
      npm install -g @anthropic-ai/claude-code
      npm install
    command: |
      echo "Claude Code ready! Run 'claude' to start."

vscode:
  extensions:
    - anthropic.claude-code

Dockerfile (.gitpod.Dockerfile)

FROM gitpod/workspace-node-lts

RUN npm install -g @anthropic-ai/claude-code

Cloud Platform Comparison

Platform	Free Tier	Persistent Storage	Best For
AWS EC2	t2.micro (750 hrs/mo)	Yes (EBS)	Full control, production-like environments
Google Cloud Shell	Free (with limits)	5GB home directory	Quick tasks, GCP-integrated projects
Azure VM	B1s (750 hrs/mo)	Yes (managed disk)	Enterprise, Azure-integrated projects
GitHub Codespaces	60 hrs/mo (2-core)	Yes (per codespace)	GitHub-hosted projects, team onboarding
Gitpod	50 hrs/mo	Limited	Open source projects, quick iterations
Railway	$5 free credit	Yes	Quick deployments, small projects

Cost Optimization

Running Claude Code in the cloud adds both compute costs and API costs. Here are strategies to minimize expenses:

Right-size instances: Claude Code itself needs minimal compute (1 vCPU, 1GB RAM). Don't overprovision.
Use spot/preemptible instances: Save 60-90% on compute for development workloads.
Auto-shutdown: Configure instances to stop after idle periods.
Use -p mode: One-shot commands are cheaper than long interactive sessions (less token overhead).
Cache results: Store Claude's analysis output to avoid re-running the same queries.
Batch operations: Group related tasks into a single Claude session rather than multiple separate calls.

Bash

# Auto-shutdown script for EC2 (add to cron)
# Stops instance after 30 minutes of no SSH sessions
#!/bin/bash
IDLE=$(who | wc -l)
if [ "$IDLE" -eq 0 ]; then
    sudo shutdown -h now
fi

Multi-Environment Setup

Configure different Claude Code settings for different environments:

Bash

# ~/.bashrc - environment-specific configuration
case "$(hostname)" in
    dev-*)
        export ANTHROPIC_API_KEY=$(aws secretsmanager get-secret-value --secret-id claude-dev --query SecretString --output text)
        ;;
    staging-*)
        export ANTHROPIC_API_KEY=$(aws secretsmanager get-secret-value --secret-id claude-staging --query SecretString --output text)
        ;;
    prod-*)
        # Read-only analysis only in production
        export ANTHROPIC_API_KEY=$(aws secretsmanager get-secret-value --secret-id claude-prod --query SecretString --output text)
        alias claude='claude --allowedTools "Read,Grep,Glob"'
        ;;
esac

Try It Yourself

Try setting up Claude Code in a GitHub Codespace — it's the quickest way to get started with cloud development. Next, learn best practices for secure and efficient remote usage.

Next: Best Practices →

← CI/CD Pipelines Best Practices →