Part of Forge DevKit ecosystem
◇ forge-product
Business language → AI-consumable specs
The problem
PRDs that AI can't parse
Product docs in Notion/Figma/Miro are designed for humans. AI agents can't extract actionable requirements.
Features built without acceptance criteria
AI jumps straight to code. No user flows, no edge cases, no definition of done.
Specs that don't match architecture
Product specs reference components that don't exist or ignore architectural boundaries.
How it works
Setup
Product auditor detects your product type: B2C, B2B, marketplace, internal tool.
Design
Describe a feature in business language. forge-product generates user flows, stories, and UX criteria.
Generate
Produces AI-consumable artifacts: user stories with AC, use cases, UX criteria matrix.
Integrate
forge-qa uses these artifacts to generate requirement-traced tests. forge-core references them in guardrails.
Key capabilities
◇9 product archetypes
B2C app, B2B SaaS, marketplace, internal tool, API service, and more. Auto-detected.
◇9 methodology frameworks
Jobs to Be Done, Impact Mapping, Story Mapping, Event Storming, and more.
◇User flows with edge cases
Happy paths, error states, and boundary conditions - all generated from feature description.
◇AI Enhancement Mixin
Augments any product archetype with AI-specific considerations: prompt design, model selection, fallbacks.
Sample output
A real-world example of what this module produces.
◆ User Story: Shopping Cart Checkout
As a registered customer on shoply-api
I want to complete checkout with saved payment methods
So that I can purchase items without re-entering card details
Acceptance Criteria:
AC-1: Saved cards displayed with last 4 digits + expiry
AC-2: New card option with Stripe Elements integration
AC-3: Order confirmation email within 30s of payment
AC-4: Cart cleared and inventory decremented atomically
Complexity: M (3-5 files) | Priority: P1Who is this for
Product Manager
Turn business requirements into structured specs that AI agents can actually execute.
Tech Lead
Translate product requirements into architecture-aware artifacts without manual spec writing.
Founder
Validate feature ideas with structured user flows before committing engineering time.
forge-product vs Figma/Miro PRD tools
See the full comparison for details.
| Dimension | Figma/Miro PRD tools | Forge DevKit |
|---|---|---|
| Input | Visual boards, manual PRDs | Plain text feature description |
| Output | Human-readable docs | AI-consumable artifacts with AC |
| Test connection | Manual handoff | Direct pipeline to forge-qa |