Getting Started

Five Steps. Then You're Building.

This section walks you through everything between "What is fai?" and "I'm productive." Five pages, read in order, and you'll understand what you have and why it matters.

Step 1: Installation

Two commands. No config files, no API keys upfront. Deno and your preferred AI tool are all you need. The whole process takes less time than reading this paragraph.

Step 2: Your First Session

What actually happens when you type fai. Real CLI output, a real conversation where you build something, and what your workbench quietly does in the background while you work.

Step 3: Your Workbench

The thing that makes fai different from every other tool: your workbench. It's not a folder. It's not a config file. It's the space where your professional identity lives - and it goes where you go.

Step 4: fai + Your AI Tool

fai doesn't replace Cursor, Claude Code, or Copilot. It sits on top of them. This page shows what each tool looks like alone versus what it looks like with fai - and why the gap widens every session after the first.

Step 5: Local Models →

Using Ollama, LM Studio, or a cloud API key instead of Claude? This page covers detection, configuration, and the synthesis priority cascade.

After Getting Started

Once you've run your first session, you'll naturally want to know:

• How does my AI remember? → How intelligence grows
• What happens over months? → How intelligence grows
• Does this work outside software? → Examples