runpod_simple

Template for simple setup of ollama + openwebUI for glm-4.7-flash

Automation CLI

Automated deployment tool for RunPod pods with SSH tunneling.

Setup

1. Copy .env.example to .env and fill in your API key:

cp .env.example .env
# Edit .env with your RUNPOD_API_KEY

2. Enter Nix shell (for NixOS) or install dependencies:

# NixOS
nix-shell

# Other systems
pip install -r requirements.txt

Usage

python main.py deploy                # Deploy pod with interactive prompts
python main.py deploy --no-reuse      # Always create new pod (skip reuse)
python main.py deploy --no-cleanup    # Keep pod running after Ctrl+C
python main.py list                  # List all pods
python main.py delete <pod_id>       # Delete specific pod

Deploy Command Options

python main.py deploy [OPTIONS]

Options:
  --no-reuse          Skip existing pod selection, always create new pod
  --template-id ID     Use specific template ID (skip selection)
  --volume-id ID       Use specific network volume ID (skip selection)
  --auto-select-gpu    Auto-select cheapest GPU without prompting
  --defaults          Use default configuration from .env (no interactive prompts)
  --no-cleanup        Don't delete pod on exit (keep running)

Notes:

• When only one template/volume exists, it's auto-selected
• With --auto-select-gpu, the cheapest GPU is chosen automatically
• With --defaults, configuration is loaded from .env without prompts
• GPU selection shows all 24GB+ options with cheapest marked with ★

Notes:

• When only one template/volume exists, it's auto-selected
• With --auto-select-gpu, the cheapest GPU is chosen automatically
• GPU selection shows all 24GB+ options with cheapest marked with ★

Features

• Interactive template and network volume selection
• Automatic GPU selection (24GB+ vRAM, cheapest option)
• Interactive GPU selection with cheapest as default
• SSH key authentication preferred, password fallback
• SSH tunnels for Ollama (11434) and WebUI (8080)
• Network-wide access (binds to 0.0.0.0) with fallback to localhost
• Automatic pod reuse for existing running pods
• Automatic pod cleanup on exit (unless --no-cleanup)
• Progress bars for pod deployment
• Clear error messages for deployment failures
• Datacenter compatibility checks
• Default configuration for unattended deployment

Configuration

For detailed information about default configuration and unattended deployment, see Defaults Documentation.

Network Access

The tool tries to bind SSH tunnels to 0.0.0.0 for network-wide access. If this fails (e.g., permission denied), it falls back to 127.0.0.1 (localhost only).

To enable network-wide access on Linux:

sudo python main.py deploy

Or configure SSH to allow user port binding in /etc/ssh/ssh_config.

Manual tunnel

For manual SSH tunneling:

ssh -4 -L 0.0.0.0:11434:127.0.0.1:11434 root@<IP> -p <PORT>
ssh -4 -L 0.0.0.0:8080:127.0.0.1:8080 root@<IP> -p <PORT>