From zero to governed in 5 minutes

One command installs the framework. Type /chati and the first agent starts — no configuration, no manual wiring.

Prerequisites

AI CLI Provider (choose one)

chati.dev works with any of these. Install at least one:

1

Initialize in your project

npx chati-dev init

Drops a chati.dev/ folder into your project with the agent definitions, governance rules, and config.

2

Activate your license

npx chati-dev activate --key=YOUR-LICENSE-KEY

Paste the license key from your dashboard. Without it, the orchestrator won't start.

3

Log in to your AI CLI provider

Authenticate with your CLI provider before starting. See your provider's documentation for the login command.

4

Start the orchestrator

Claude Code / Gemini CLI
/chati
OpenAI Codex
$chati

The orchestrator reads your project, detects whether it's a new build or an existing codebase, and hands you off to the first agent.

Updating

Already installed? Update without losing your project artifacts.

npx chati-dev update

Updates constitution, agents, and hooks to latest. Your project artifacts (PRD, tasks, handoffs) are preserved.

What happens next?

The pipeline guides you through:

  1. DISCOVER: Capture your goals and the current state of the project
  2. PLAN: Turn requirements into architecture, design, phases, and tasks — with a 95% quality gate before moving on
  3. BUILD: Implement following the plan, with automated code and visual quality gates
  4. DEPLOY: Ship to production with CI/CD, security checks, and rollback paths

Every step leaves a paper trail, every decision is logged, and nothing ships without passing a quality check.