This project was built on Windows, but Ollama works great on Mac and Linux too.
# Download from ollama.com or use Homebrew:
brew install ollama# Find your USB drive (usually /Volumes/USB321FD or similar)
export OLLAMA_MODELS="/Volumes/YOUR_USB/ollama/models"
# Start server
ollama serve &
# Pull the model
ollama pull dolphin-llama3
# Run it
ollama run dolphin-llama3Save this as start.sh on your USB root:
#!/bin/bash
DRIVE=$(dirname "$(realpath "$0")")
export OLLAMA_MODELS="$DRIVE/ollama/models"
echo "AI Survival USB - Starting..."
echo "Drive: $DRIVE"
echo "Models: $OLLAMA_MODELS"
ollama serve &
sleep 5
ollama run dolphin-llama3Make it executable: chmod +x start.sh
Note: On Mac, you'll need Ollama installed on the system (via brew or the .dmg). The USB carries the model, but the Ollama binary runs from the Mac.
curl -fsSL https://ollama.com/install.sh | sh# Find your USB (check with lsblk or mount)
# Usually /media/username/USB321FD or /mnt/usb
export OLLAMA_MODELS="/media/$USER/YOUR_USB/ollama/models"
# Start and pull
ollama serve &
sleep 5
ollama pull dolphin-llama3
# Run
ollama run dolphin-llama3Same as the Mac version above. Save to USB root, chmod +x start.sh.
For a truly portable Linux setup, you can include the Ollama binary on the USB:
# Download the Linux binary
curl -L https://ollama.com/download/ollama-linux-amd64 -o /media/$USER/YOUR_USB/ollama-linux
chmod +x /media/$USER/YOUR_USB/ollama-linuxThen your start script can reference the binary directly from the USB without needing Ollama pre-installed on the system.
- No
.batfiles — use.shscripts instead - File paths use
/not\ - USB drives mount at
/Volumes/(Mac) or/media/(Linux) instead of drive letters - Everything else is identical — same models, same commands, same results