Google Cloud Free Tier

Google Cloud's always-free e2-micro VM gives you a permanent server with 1GB RAM and 30GB disk — enough for base Crow. Chain it with an Oracle Cloud instance for redundant, multi-cloud Crow.

Why Google Cloud as a secondary?

• Always free — e2-micro never expires, no surprise charges
• 1 GB RAM, 30 GB disk — runs base Crow comfortably
• Chains with Oracle Cloud — two free clouds, synced automatically
• US regions — us-west1, us-central1, or us-east1 (free tier restriction)

Step 1: Create a Google Cloud Account

1. Go to cloud.google.com and click Get started for free
2. Sign in with your Google account
3. Enter billing information — Google requires a credit card for verification. The e2-micro instance is genuinely free and you will not be charged as long as you stay within the free tier limits.
4. You'll receive $300 in free trial credits (valid 90 days) but do not need them — the e2-micro is in the Always Free tier, separate from the trial.

Step 2: Create an e2-micro VM

1. Go to the Google Cloud Console
2. Navigate to Compute Engine → VM Instances (enable the API if prompted)
3. Click Create Instance
4. Configure:
   • Name: crow
   • Region: us-central1 (or us-west1, us-east1 — only these three regions are free)
   • Zone: Any available zone in your chosen region
   • Machine type: Under General purpose → E2, select e2-micro (0.25 vCPU, 1 GB RAM)
     • Look for "1 free e2-micro instance per month" in the Always Free tier docs
   • Boot disk: Click Change → Ubuntu 22.04 LTS → Size: 30 GB → Standard persistent disk
   • Firewall: Do NOT check "Allow HTTP traffic" or "Allow HTTPS traffic" — we'll use Tailscale for private access
5. Click Create

The VM takes about a minute to start. Note the External IP on the instances list.

Free tier limits

The e2-micro free tier includes 1 instance, 30GB disk, and 1GB egress/month to regions outside North America. Exceeding these limits incurs charges. Monitor your billing dashboard.

Step 3: Connect via SSH

Click the SSH button next to your VM in the Console, or from your local machine:

# Add your SSH key to the VM via the Console:
# Compute Engine → Metadata → SSH Keys → Add SSH Key
ssh your-username@<EXTERNAL_IP>

Step 4: Install Crow

# Install Node.js 20
curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt-get install -y nodejs git

# Clone and set up Crow
git clone https://github.com/kh0pper/crow.git ~/crow
cd ~/crow
npm install
npm run setup
npm run init-db

Verify:

node servers/memory/index.js
# Should start without errors — press Ctrl+C to stop

Step 5: Security Hardening

# Enable firewall — only allow SSH and Tailscale
sudo ufw default deny incoming
sudo ufw default allow outgoing
sudo ufw allow ssh
sudo ufw allow 41641/udp  # Tailscale
sudo ufw enable

# Install fail2ban
sudo apt install -y fail2ban
sudo systemctl enable fail2ban

# Disable SSH password auth (use keys only)
sudo sed -i 's/#PasswordAuthentication yes/PasswordAuthentication no/' /etc/ssh/sshd_config
sudo sed -i 's/PasswordAuthentication yes/PasswordAuthentication no/' /etc/ssh/sshd_config
sudo systemctl restart sshd

# Enable automatic security updates
sudo apt install -y unattended-upgrades
sudo dpkg-reconfigure -plow unattended-upgrades

Also remove the default GCP firewall rules that allow HTTP/HTTPS from anywhere:

1. Go to VPC Network → Firewall in the Google Cloud Console
2. Find default-allow-http and default-allow-https
3. Delete them (or set them to disabled) — Tailscale handles all access

Step 6: Install Tailscale

curl -fsSL https://tailscale.com/install.sh | sh
sudo tailscale up

Follow the authentication URL. Once connected:

# Note your Tailscale IP
tailscale ip -4

Allow Crow's gateway port through UFW for Tailscale traffic:

# Allow port 3001 from Tailscale network only
sudo ufw allow from 100.64.0.0/10 to any port 3001

Step 7: Start the Gateway

cd ~/crow

# Create a systemd service for persistence
# Replace YOUR_USERNAME with your actual username (run `whoami` to check)
sudo tee /etc/systemd/system/crow-gateway.service > /dev/null << 'EOF'
[Unit]
Description=Crow Gateway
After=network.target

[Service]
Type=simple
User=YOUR_USERNAME
WorkingDirectory=/home/YOUR_USERNAME/crow
ExecStart=/usr/bin/node servers/gateway/index.js --no-auth
Restart=always
RestartSec=5
Environment=PORT=3001

[Install]
WantedBy=multi-user.target
EOF

sudo systemctl daemon-reload
sudo systemctl enable crow-gateway
sudo systemctl start crow-gateway

Verify:

sudo systemctl status crow-gateway
curl http://localhost:3001/health

Step 8: Connect Your AI Platform

Your Crow server is now running and accessible via Tailscale. Connect it from any AI platform:

• Claude Web & Mobile — http://<tailscale-ip>:3001/memory/mcp
• ChatGPT — http://<tailscale-ip>:3001/memory/sse
• Gemini — http://<tailscale-ip>:3001/memory/mcp
• Claude Code — http://<tailscale-ip>:3001/memory/mcp
• All platforms

Visit http://<tailscale-ip>:3001/setup from a device on your Tailscale network to see integration status and endpoint URLs.

Try it out

After connecting your AI platform, say:

"Remember that today is my first day using Crow" "What do you remember?"

Step 9: Chain with Oracle Cloud

If you have an Oracle Cloud instance running Crow, you can chain them together so memories sync automatically and you can call tools on either from either.

Export identity from Oracle Cloud

On your Oracle Cloud instance (the home instance):

cd ~/crow
npm run identity:export

Copy the export file to your Google Cloud instance:

scp ~/.crow/identity-export.enc your-username@<google-cloud-external-ip>:~/

Import identity on Google Cloud

On Google Cloud:

cd ~/crow
npm run identity:import
# Enter the same passphrase used during export

Verify both have the same Crow ID:

npm run identity
# Should show the same crow:xxxxxxxxxx on both machines

Register instances

Tell your AI on the Oracle Cloud instance:

"Register my Google Cloud instance as a satellite at http://<google-tailscale-ip>:3001,
hostname google-cloud, name Cloud Satellite"

Tell your AI on the Google Cloud instance:

"Register my Oracle Cloud instance as home at http://<oracle-tailscale-ip>:3001,
hostname oracle-cloud, name Oracle Home"

Verify federation

On either machine:

"List instances"

You should see both listed. Test sync:

# On Oracle Cloud
"Remember that my Google Cloud satellite is working"

# On Google Cloud (wait a moment)
"Search memories for Google Cloud satellite"

The memory should appear on both instances.

What you get from chaining

• Redundancy — if one cloud goes down, the other has your data
• Federation — call tools on Oracle from Google Cloud and vice versa
• Free tier stacking — Oracle (1GB RAM) + Google Cloud (1GB RAM) = more capacity for separate workloads

For the full multi-device reference, see Multi-Device Quick Start.

Optional: Make Your Blog Public

For public blog setup, see the Oracle Cloud guide's blog section — the steps are identical for Google Cloud.

What to Do If Compromised

1. Stop the instance — Compute Engine → VM Instances → Stop
2. Backup your data — if you can still SSH in: cd ~/crow && npm run backup
3. Rotate SSH keys — delete old keys from Console Metadata, add new ones
4. Check logs — sudo grep "Failed password" /var/log/auth.log | tail -20
5. Re-image if needed — delete and recreate the VM. Restore data with npm run restore