Unified AI Development Framework - BMAD phases with Ralph execution loop. Bmalph integrates BMAD-METHOD for AI-driven planning (Phases 1-3) with the Ralph autonomous implementation loop (Phase 4), streamlining the AI development lifecycle from concept to code. It targets engineers and power users seeking a structured, automated approach to software development using AI coding assistants.

How It Works

Bmalph bundles two core AI development systems: BMAD-METHOD for planning and Ralph for autonomous implementation. BMAD guides users through analysis, planning, and solutioning phases using AI agents and workflows tailored to specific platforms like Claude Code, OpenAI Codex, and Cursor. Once planning artifacts are generated, bmalph implement transitions them to a format Ralph understands. Ralph then executes an autonomous loop, picking up tasks, implementing them with Test-Driven Development (TDD), committing changes, and iterating until completion, managed via platform-specific drivers.

Quick Start & Requirements

• Primary install: npm install -g bmalph
• Prerequisites: Node.js 20+, Bash (WSL or Git Bash on Windows), and a supported AI coding platform (e.g., Claude Code, OpenAI Codex, Cursor). Ralph (Phase 4) requires a "full tier" platform with CLI support.
• Setup: Navigate to your project directory and run bmalph init --name <your-project-name>. The --platform flag can specify the AI assistant.
• Links: Official quick-start is via CLI commands; see CONTRIBUTING.md for development setup.

Highlighted Details

• Supports multiple AI coding assistants with distinct tiers: "full" (Phases 1-4) and "instructions-only" (Phases 1-3).
• Phased workflow: Analysis (BP, MR, DR, TR, CB), Planning (CP, VP, EP, CU), Solutioning (CA, CE, IR), and Implementation (Ralph).
• Ralph's autonomous loop features TDD, automatic commits, a circuit breaker for infinite loops, and a response analyzer.
• Incremental development is supported, allowing Ralph to pick up new epics after initial implementation.
• "Smart Merge" preserves completed stories and integrates new ones when bmalph implement is re-run.

Maintenance & Community

No specific community channels (Discord/Slack) or prominent contributors are detailed in the README. Development guidelines are available in CONTRIBUTING.md.

Licensing & Compatibility

• License: MIT.
• Compatibility: The MIT license generally permits commercial use and integration with closed-source projects.

Limitations & Caveats

Support for GitHub Copilot and Cursor is experimental. Platforms designated "instructions-only" (Windsurf, Aider) do not support the Ralph autonomous implementation loop (Phase 4). Windows users require Git Bash or WSL for proper Bash execution. The bmalph watch command is deprecated.