What is BMAD-METHOD?

BMAD-METHOD™ (Breakthrough Method for Agile AI Driven Development) is an open-source universal AI agent framework that transforms how we approach AI-driven development across multiple domains.

Unlike traditional AI coding assistants, BMAD introduces a collaborative multi-agent system where specialized AI agents work together to handle the entire development lifecycle - from planning to deployment.

Core Innovation

BMAD-METHOD addresses two critical challenges in AI-assisted development:

1. Agentic Planning

Specialized AI agents (Analyst, Project Manager, Architect) collaborate to create detailed product requirements and architecture documents using advanced prompt engineering and human-in-the-loop refinement.

2. Context-Engineered Development

The Scrum Master agent transforms planning documents into hyper-detailed development stories that embed full context, implementation details, and architectural guidance - eliminating “planning inconsistency” and “context loss.”

Agent Collaboration Workflow

graph LR
    A[Analyst Agent] --> PM[Project Manager]
    PM --> AR[Architect Agent]
    AR --> SM[Scrum Master]
    SM --> DEV[Development Agent]
    DEV --> QA[QA Agent]
    QA --> |Feedback| SM

    style A fill:#e1f5fe
    style PM fill:#f3e5f5
    style AR fill:#e8f5e9
    style SM fill:#fff3e0
    style DEV fill:#fce4ec
    style QA fill:#f1f8e9

The Six AI Agents

flowchart TD
    subgraph Planning["🎯 Planning Phase"]
        A[📊 Analyst Agent<br/>Requirements Analysis]
        PM[📋 Project Manager<br/>Project Planning]
        AR[🏗️ Architect Agent<br/>System Design]
    end

    subgraph Execution["⚙️ Execution Phase"]
        SM[📝 Scrum Master<br/>Story Creation]
        DEV[💻 Development Agent<br/>Code Implementation]
        QA[✅ QA Agent<br/>Quality Assurance]
    end

    A --> PM --> AR
    AR --> SM --> DEV --> QA

    style Planning fill:#e3f2fd
    style Execution fill:#f3e5f5

Agent Responsibilities

AgentRoleKey Outputs
AnalystGathers and analyzes requirementsDetailed PRD (Product Requirements Document)
Project ManagerPlans project scope and timelineProject plan, milestones
ArchitectDesigns system architectureArchitecture diagrams, technical specs
Scrum MasterCreates development storiesContext-rich user stories
DeveloperImplements featuresWorking code, documentation
QATests and validatesTest results, bug reports

Key Features

  • Domain Agnostic: Works across software development, creative writing, business strategy, education, and more
  • 🔌 Expansion Packs: Domain-specific AI agents available through expansion packs
  • 🤝 Collaborative: Agents work together, maintaining context throughout the process
  • 📚 Context Preservation: Eliminates context loss between planning and development
  • 🔄 Human-in-the-Loop: Refinement at every stage ensures quality
  • 🌐 Open Source: MIT License, Node.js-based

How It Works

sequenceDiagram
    participant User
    participant Analyst
    participant PM
    participant Architect
    participant ScrumMaster
    participant Developer

    User->>Analyst: Project Idea
    Analyst->>PM: Requirements Doc
    PM->>Architect: Project Plan
    Architect->>ScrumMaster: Architecture Specs
    ScrumMaster->>Developer: Context-Rich Stories
    Developer->>User: Implemented Features

Use Cases

BMAD-METHOD excels in:

  • 🚀 Software Development: Full-stack applications, APIs, microservices
  • ✍️ Creative Writing: Novels, scripts, content creation
  • 💼 Business Strategy: Business plans, market analysis
  • 🎓 Education: Curriculum development, learning materials
  • 🎮 Entertainment: Game design, interactive experiences
  • 💪 Personal Wellness: Health programs, fitness plans

Getting Started

  1. Visit the GitHub Repository: github.com/bmad-code-org/BMAD-METHOD
  2. Join the Community: Connect on Discord for support
  3. Explore Expansion Packs: Find domain-specific agents for your needs
  4. Start Building: Follow the documentation to create your first BMAD project

Why BMAD Matters

Traditional AI coding assistants often struggle with:

  • Losing context between conversations
  • Inconsistent planning and implementation
  • Lack of domain expertise
  • Poor collaboration between different development phases

BMAD-METHOD solves these problems by providing a structured, multi-agent approach that maintains context, ensures consistency, and delivers comprehensive solutions from inception to deployment.


Learn More: BMAD-METHOD on GitHub

BMAD-METHOD is open-source software licensed under MIT License.