Getting Started

PAI boots from a USB drive into a complete, private AI workstation. Nothing is installed on your computer — everything runs in RAM.

Requirements

USB drive: 8 GB minimum (32 GB+ recommended for AI models)
Target machine: x86_64 (Intel or AMD, 64-bit)
RAM: 4 GB minimum — 8 GB recommended for 7B parameter models
Boot order: ability to select USB from BIOS/UEFI (F12 / F2 / DEL at POST)

Step 1: Flash the ISO

Linux / macOS (one command):

curl -fsSL https://raw.githubusercontent.com/nirholas/pai/main/scripts/flash.sh | sudo bash

The script detects USB drives, shows your options, and asks for confirmation before writing.

Manual (Linux):

# Find your USB device
lsblk -d -o NAME,SIZE,MODEL,TRAN | grep usb

# Flash
sudo dd if=pai.iso of=/dev/sdX bs=4M status=progress && sync

Windows: Download the ISO and use Rufus — select your USB, select the ISO, click Start.

Step 2: Boot

Plug the USB into your target machine
Reboot and press F12 (or F2/DEL/ESC) at the POST screen
Select your USB drive from the boot menu
PAI loads into RAM — usually takes 15–30 seconds

Step 3: Talk to AI

Firefox opens automatically to the PAI chat interface at localhost:8080.

Click Models → pull a model (e.g., llama3.2 for a fast 3B model)
Wait for the download (only needed once per session)
Start chatting — your prompts never leave your machine

# Or use the terminal (Alt+Enter to open):
ollama pull llama3.2
ollama run llama3.2 "Explain how PAI protects my privacy"

Keyboard shortcuts

Shortcut	Action
`Alt + Enter`	Open terminal
`Alt + D`	Application launcher
`Alt + F4`	Close window
`Alt + L`	Lock screen
`Alt + 1–9`	Switch workspace

Privacy commands

# Enable Tor (routes all traffic through Tor)
sudo pai-privacy on

# Disable Tor
sudo pai-privacy off

# Check your randomized MAC address
ip link show | grep ether

Shutdown

Power off normally. When PAI shuts down, all RAM is cleared. Your chat history, downloaded files, and any credentials are gone — the machine returns to exactly how it was before you booted PAI.