BMAD Method¶
BMAD stands for Build More Architect Dreams — an AI-driven agile development framework that guides teams from rough idea to working code through structured, phase-gated workflows.
Core Philosophy¶
Traditional AI tools do the thinking for you, producing average results. BMAD agents and facilitated workflows act as expert collaborators who guide you through a structured process to bring out your best thinking in partnership with AI.
The framework is built on a key insight: AI agents work best with clear, structured context. BMAD builds that context progressively across 4 distinct phases, where each phase and workflow produces documents that inform the next — so agents always know what to build and why.
The 4 Phases¶
| Phase | Name | What Happens |
|---|---|---|
| 1 | Analysis | Brainstorming, research, product brief or PRFAQ |
| 2 | Planning | Create requirements (PRD or spec) |
| 3 | Solutioning | Design architecture (BMad Method/Enterprise only) |
| 4 | Implementation | Build epic by epic, story by story |
The 9-Section PRD Structure¶
At the heart of BMAD is a standardized PRD format with 9 structured sections. This structure is used by Sharpener.dev and forms the basis for comprehensive product scoping.
| # | Section | Purpose |
|---|---|---|
| 1 | Discovery | Understand what you're building and who it's for. Problem statement, target users, pain points. |
| 2 | Success Criteria | How do we measure success? Broken into User, Business, and Technical success metrics. |
| 3 | User Journeys | Map the key paths users take through the product. Happy paths AND edge cases. |
| 4 | Domain | Define the product category and its rules. Regulatory, business logic, taxonomy. |
| 5 | Innovation | What's novel about this product? Core assumptions that must be validated. |
| 6 | Product Type | Mobile app? Web platform? API? Library? Determines technical constraints. |
| 7 | Scope | Define MVP vs. post-MVP features. Make explicit tradeoffs about what to defer. |
| 8 | Functional Requirements | Concrete list of what the system must do. |
| 9 | Non-Functional Requirements | Performance, security, accessibility, scalability, SLAs. |
Why 9 Sections?¶
The 9-section structure is designed to surface what you don't know vs. what you think you know. Each section has dedicated elicitation techniques (50+ per section) that probe for gaps, edge cases, and assumptions.
The goal is not to fill out a template — it's to force systematic pressure-testing of every aspect of a product idea before building begins.
Elicitation Techniques¶
BMAD provides 50+ elicitation techniques for going deeper on each section. Examples include:
- Pain-point mapping — For Discovery: what specific pain does this solve and how acute is it?
- Success metric definition — For Success Criteria: how would we know if this failed? What metrics would show that?
- Journey stress-testing — For User Journeys: what happens at each step if the user is offline? On mobile with poor connectivity? Under time pressure?
- Assumption inversion — For Innovation: what must be true for this to work? What if that assumption fails?
- Tradeoff explicit budgeting — For Scope: if we had to cut 40% of features, what would we keep and why?
Planning Tracks¶
BMAD adapts to project complexity through three planning tracks:
| Track | Best For | Documents Created |
|---|---|---|
| Quick Flow | Bug fixes, simple features, clear scope (1-15 stories) | Tech-spec only |
| BMad Method | Products, platforms, complex features (10-50+ stories) | PRD + Architecture + UX |
| Enterprise | Compliance, multi-tenant systems (30+ stories) | PRD + Architecture + Security + DevOps |
Key Agents¶
BMAD includes specialized AI agents for each domain:
- PM Agent — Product management, PRD creation, epics and stories
- Architect Agent — Technical architecture and ADRs
- Developer Agent — Implementation, sprint planning, code review
- UX Agent — User experience design
- Analyst Agent — Market, domain, and technical research
- 12+ total domain experts for specialized workflows
Party Mode¶
BMAD supports "Party Mode" — bringing multiple agent personas into one session to collaborate and discuss, simulating a cross-functional team review.
Implementation with AI Coding Tools¶
BMAD is designed to work with any AI coding assistant that supports custom system prompts:
- Claude Code (recommended)
- Cursor
- Codex CLI
The framework outputs structured documents (PRD, architecture, stories) that become project context — ensuring AI agents make consistent, well-grounded decisions throughout implementation.
Installation¶
npx bmad-method install
Prerequisites: Node.js v20.12+ and Python 3.10+.
Documentation¶
- Docs: https://docs.bmad-method.org
- GitHub: https://github.com/bmad-code-org/BMAD-METHOD
- Discord: https://discord.gg/gk8jAdXWmj
Relationship to Sharpener.dev¶
Sharpener.dev is a practical front-end application of the BMAD Method. Where BMAD is the framework, Sharpener is the tool that applies it to generate a complete PRD from a single-sentence idea. The 9-section PRD structure used by Sharpener directly maps to BMAD's elicitation framework — and aligns with the Kelly factory's pipeline-stages Stage 3 (Planning) output requirements, where quality-gates validate completeness before moving to implementation.
The BMAD 9-section structure also mirrors the factory's full-pipeline approach: both enforce structured gates (BMAD's section-by-section elicitation vs. the factory's READY/NOT-READY and PASS/FAIL gates) to prevent shallow or incomplete plans from propagating into build.
See Also¶
- Sharpener.dev
- Plantry PRD Demo (example output)
- BMAD Library — agent definition library using this framework
- Factory Rules — foundational factory operating rules
- Quality Gates — validates PRD completeness in Stage 3
- Discovery Gate — 28/40 gate validates opportunity before planning; BMAD's structured elicitation improves gate pass rates
Bibliography¶
- BMAD Method Documentation — https://docs.bmad-method.org
- BMAD Method GitHub — https://github.com/bmad-code-org/BMAD-METHOD
- Sharpener.dev (application of BMAD) — https://sharpener.dev/