How to Set Up Multi-Machine GPU Orchestration in 5 Minutes

# Download and install rbee
curl -fsSL https://rbee.dev/install.sh | sh

# Verify installation
rbee --version

# Add your first machine
rbee hive add gaming-pc 192.168.1.100 \
  --user your-username \
  --ssh-key ~/.ssh/id_rsa

# Add a Mac Studio
rbee hive add mac-studio 192.168.1.101 \
  --user your-username

# Add a remote server
rbee hive add cloud-gpu ssh.example.com \
  --user ubuntu \
  --port 2222

# List all configured hives
rbee hive list

# SSH into each machine and run:
curl -fsSL https://rbee.dev/install-worker.sh | sh

# Or let rbee deploy automatically:
rbee worker deploy --all-hives

# Download Llama 3.1 8B to all hives
rbee model download llama-3.1-8b --all-hives

# Or download to specific hives
rbee model download llama-3.1-70b \
  --hive gaming-pc \
  --hive mac-studio

# List available models
rbee model list

# Start the keeper (runs on port 7833 by default)
rbee keeper start

# Or run in the background
rbee keeper start --daemon

# Check status
rbee keeper status

# Start workers on all hives
rbee worker start --all-hives

# Or start on specific hives
rbee worker start --hive gaming-pc --hive mac-studio

# Check worker status
rbee worker list

# Test with curl
curl http://localhost:7833/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "llama-3.1-8b",
    "messages": [
      {"role": "user", "content": "Hello! What can you do?"}
    ]
  }'

# Or use the rbee CLI
rbee chat "Tell me about GPU orchestration"

[keeper]
host = "0.0.0.0"
port = 7833
log_level = "info"

[[hives]]
name = "gaming-pc"
host = "192.168.1.100"
user = "your-username"
ssh_key = "~/.ssh/id_rsa"

[[hives]]
name = "mac-studio"
host = "192.168.1.101"
user = "your-username"

[routing]
strategy = "round-robin"  # or "least-loaded", "custom"