Spec Kitty Documentation

Build high-quality software faster with AI-powered spec-driven development.

What is Spec Kitty?

Spec Kitty is a toolkit for spec-driven development with AI coding agents. It structures your AI workflows around specifications, plans, and work packages—ensuring that AI agents build exactly what you need, with live progress tracking via a kanban dashboard.

Works with Claude Code, Cursor, Windsurf, Gemini CLI, GitHub Copilot, and 7 other AI coding agents.

Release Tracks

Spec Kitty runs two active release lines:

main / PyPI 1.x (recommended): install with pip install spec-kitty-cli
2.x / GitHub-only: semantic tags (v2.*.*) and release artifacts for forward-track testing

If you're new to Spec Kitty, start on the PyPI 1.x line.

📚 Tutorials — Learning-Oriented

New to Spec Kitty? Start here to learn the fundamentals.

Claude Code Integration — Setup and first run
Claude Code Workflow — End-to-end workflow walkthrough

🔧 How-To Guides — Task-Oriented

Solve specific problems with step-by-step instructions.

Install & Upgrade — Installation methods and upgrades
Use the Dashboard — Monitor progress in real-time
Run External Orchestrator — Drive multi-agent automation via spec-kitty-orchestrator
Build Custom Orchestrator — Build your own provider against the host API
Non-Interactive Init — Run init without prompts

📖 Reference — Information-Oriented

Complete command and configuration documentation.

CLI Commands — All spec-kitty commands
Orchestrator API — Contract for external orchestrators
Slash Commands — All /spec-kitty.* commands
Agent Subcommands — spec-kitty agent * commands
Configuration — Config files and options
Environment Variables — All env vars
File Structure — Directory layout
Missions — Mission types and configuration
Supported Agents — All 12 supported AI agents

💡 Explanations — Understanding-Oriented

Understand the concepts and architecture.

Spec-Driven Development — The philosophy
Workspace-per-WP Model — How workspaces work
Git Worktrees — Git worktrees explained
Mission System — Why missions exist
Kanban Workflow — Lane-based workflow
AI Agent Architecture — Multi-agent design
Multi-Agent Orchestration — Manual and external orchestration models

Dashboard Preview

Spec Kitty includes a live dashboard for real-time progress tracking:

Spec Kitty Dashboard - Kanban Board View

Kanban board showing work packages across all lanes

Feature overview with completion metrics

Quick Start

# Install
pip install spec-kitty-cli

# Initialize a new project
spec-kitty init my-project --ai claude
cd my-project

# Launch your AI agent and use slash commands
/spec-kitty.specify Add user authentication with email/password
/spec-kitty.plan
/spec-kitty.tasks
/spec-kitty.implement

Ready to start? Getting Started Tutorial →

Support

Issues: GitHub Issues
Discussions: GitHub Discussions