AI-Powered Development Environment with BMAD Methodology
Your complete AI development toolkit, running locally on your Mac
Catalyst helps developers harness AI assistants more effectively by providing a complete, privacy-focused toolkit that runs entirely on your Mac. Think of it as your AI development command center - everything you need to work with AI coding assistants, all in one simple package.
The Problem: AI coding tools are fragmented and cloud-dependent. Setting up Continue.dev, managing MCP (Model Context Protocol) servers, configuring local AI models, and following structured development methodologies requires hours of setup across dozens of tools.
The Solution: Catalyst brings everything together in one Homebrew-installable package. Install once, get instant access to 20+ integrated development tools, local AI support, and the complete BMAD (Brian Madison Development) methodology for structured software development.
Key Benefits:
- Privacy First - Everything runs on your machine; cloud APIs are completely optional
- One-Command Setup -
brew install catalystgets you a complete AI development environment - Proven Methodology - BMAD integration provides 10 AI agent roles for the complete software development lifecycle
- Extensible - Open source and designed for customization
# Install Catalyst
brew install catalyst
# Run the interactive setup wizard
catalyst setup
# Initialize in your project
cd your-project
catalyst init
# Verify everything works
catalyst verifyExpected output:
✅ Catalyst 0.1.0
✅ BMAD Method installed
✅ MCP servers configured
✅ All systems operational!
Next Steps: Check out the User Guide for detailed usage instructions and tutorials.
- BMAD Methodology Integration - Complete structured development framework with 10 AI agent roles (Product Owner, Developer, QA, Architect, and more)
- 20+ MCP Server Management - GitHub, Docker, PostgreSQL, Git, Filesystem, Browser automation, and more - all configured automatically
- LM Studio Integration - Run AI models locally with MLX optimization for Apple Silicon (M1/M2/M3)
- Development Automation - Structured workflows, templates, and tasks for the complete software development lifecycle
- Multi-IDE Support - Works with VS Code, Cursor, Claude Code, Windsurf, and 15+ other IDEs
- Open-Source & Extensible - AGPL v3.0 licensed, fully customizable and transparent
# Example: Using Catalyst with BMAD agents in your IDE
# Initialize a new project with BMAD methodology
catalyst init --type fullstack
# The Product Owner agent creates structured requirements
@po create epic for user authentication
# Developer agent implements the story with tests
@dev implement story 1.2 from epic-1-auth.md
# QA agent reviews implementation and runs tests
@qa review story 1.2
# All coordinated through MCP servers - Git, GitHub, Docker, etc.How it works: Catalyst configures your IDE with MCP servers that give AI assistants real capabilities - creating GitHub PRs, running Docker containers, querying databases, and more. Combined with BMAD methodology, you get structured, professional development workflows powered by AI.
Perfect for developers who want to use Catalyst in their daily work:
- User Guide - Complete installation and usage guide
- Development Guide - Local development workflow and best practices
- LM Studio Setup - Configure local AI models
- Troubleshooting - Common issues and solutions
For those who want to understand how Catalyst works under the hood:
- Architecture Overview - System design and technical details
- Technology Stack - Technologies and design decisions
- CLI Architecture - Command-line interface design
- MCP Server Architecture - How MCP servers are managed
- Project Structure - Codebase organization
Ready to help make Catalyst better?
- Contributing Guide - How to contribute code, documentation, and ideas
- Development Setup - Set up local development environment
- Code of Conduct - Community standards and expectations
- Good First Issues - Great starting points for new contributors
Join the Catalyst community to get help, share ideas, and collaborate:
- GitHub Discussions - Ask questions, share use cases, discuss ideas
- Issue Tracker - Report bugs and request features
- Security Reports - Responsible disclosure of security vulnerabilities
Getting Help:
- Check the User Guide and Troubleshooting
- Run
catalyst doctorto diagnose common issues - Search GitHub Discussions for similar questions
- Create a new discussion or issue if you need help
Code of Conduct: We are committed to providing a welcoming and inclusive community. Please read our Code of Conduct.
Catalyst includes these MCP servers, configured automatically:
Essential (7 servers):
- GitHub - Repository management, issues, pull requests, workflows
- Git - Local repository operations
- Filesystem - Safe file operations with permissions
- Docker - Container management and orchestration
- PostgreSQL - Database operations and queries
- Fetch - Web content retrieval and HTTP requests
- Memory - Persistent knowledge graph for AI context
Development Tools (5+ servers):
- Vite - Frontend build and development server
- Xcode - iOS and macOS development tools
- Storybook - Component documentation and testing
- Playwright - Browser automation and end-to-end testing
- Figma - Design-to-code integration
All servers are pre-configured and work immediately after catalyst setup. Learn more in the MCP Server Architecture docs.
You: "Hey AI, create a GitHub PR for this bug fix"
AI: "Here are the git commands you should run..."
You: *copies commands, runs manually, switches to browser*
You: *manually creates PR in GitHub UI*
You: *switches back to IDE, forgets what AI suggested*
You: "@dev Fix the authentication bug, write tests, and open a PR"
AI: *analyzes code using Filesystem MCP*
AI: *implements fix and generates tests*
AI: *runs tests using local development tools*
AI: *commits changes using Git MCP*
AI: *creates PR using GitHub MCP*
Result: "✅ PR #123 created with all tests passing. Ready for review."
The Difference: Catalyst gives AI assistants real capabilities through MCP servers, combined with structured methodology through BMAD. The result is AI that actually does the work, not just suggests it.
Required:
- macOS 13.0+ (Ventura or later)
- Homebrew package manager
- 16GB+ RAM recommended
- 50GB+ free disk space
Optional but Recommended:
- Continue.dev or Claude Code extension (automatically detected during setup)
- LM Studio for local AI models
- Docker Desktop (for Docker MCP server)
- PostgreSQL (for database MCP server)
Note: Catalyst automatically detects your installed IDEs and tools during catalyst setup and configures them appropriately.
# Setup and initialization
catalyst setup # Interactive setup wizard
catalyst init # Initialize project with BMAD
catalyst verify # Verify installation status
catalyst update # Update BMAD templates and config
catalyst doctor # Troubleshoot common issues
# MCP server management
catalyst mcp list # List available MCP servers
catalyst mcp test <name> # Test specific server
catalyst mcp logs <name> # View server logs
# Configuration
catalyst config show # Show current configuration
catalyst config edit # Edit configuration fileFull command documentation: CLI Architecture
Catalyst is licensed under the GNU Affero General Public License v3.0 (AGPL v3.0).
What this means:
- ✅ Free and open source forever - Use Catalyst at no cost
- ✅ Modify and distribute freely - Fork, customize, and share your changes
- ✅ Share improvements - If you modify and deploy Catalyst (even as a service), you must share your changes
- ✅ Prevents commercial exploitation - Forks cannot be commercialized without giving back to the community
Why AGPL? We chose AGPL v3.0 to ensure Catalyst remains free and open source, while preventing commercial entities from taking the code without contributing improvements back to the community.
See the LICENSE file for complete details.
Catalyst stands on the shoulders of giants:
- BMAD-METHOD (Brian Madison) - Structured development methodology and agent framework
- Anthropic - Model Context Protocol (MCP) specification and tooling
- LM Studio - MLX-optimized local AI model hosting for Apple Silicon
- Continue.dev - Excellent open-source IDE integration
- Homebrew - macOS package distribution platform
- All MCP server contributors - Open-source community building the MCP ecosystem
Current Version: 0.1.0 (MVP - Active Development)
Completed:
- ✅ Core CLI implementation
- ✅ Essential MCP servers (7 servers)
- ✅ BMAD methodology integration
- ✅ Development documentation
- ✅ Open-source governance
In Progress:
- 🔄 Build and release automation (Epic 5)
- 🔄 Comprehensive user documentation (Epic 6)
Planned:
- 📋 Homebrew formula and public distribution
- 📋 Additional MCP servers (Redis, MongoDB, Jenkins)
- 📋 Enhanced IDE detection and configuration
- 📋 Community-built MCP server marketplace