SELF-INITIATED OPEN-SOURCE PROJECT • AI PRODUCT • REQUIREMENTS INTELLIGENCE

From feature idea to engineering-ready requirements


Forge is an AI-powered product that transforms natural-language feature ideas into structured engineering artifacts including user stories, acceptance criteria, API contracts, workflow definitions, and Jira-ready tickets.

I built Forge to solve a common product delivery challenge: requirements often appear complete in prose but hide missing assumptions, dependencies, edge cases, and technical constraints until development begins.

Instead of acting as another AI writing tool, Forge guides users through structured context collection before generating delivery-ready outputs with built-in confidence indicators.

<aside>

Project Scope

Forge is a self-initiated open-source portfolio project. I owned the product strategy, workflow design, prompt architecture, user testing, and product decisions. Claude powers the AI generation layer, while all product thinking and system design are my own.

</aside>

Project Snapshot


<aside>

3 Workflow Stages

</aside>

<aside>

5 Generated Artifacts

</aside>

<aside>

18 Early Test Users

</aside>

<aside>

127 Prompt Iterations

</aside>

Summary


Problem

Requirements are often incomplete and inconsistent before engineering begins.

Opportunity

Help Product Managers produce complete, structured, review-ready requirements using AI.

Solution

An AI-powered workflow that converts feature descriptions into engineering-ready artifacts while exposing uncertainty through confidence flags.

The Problem


Traditional requirement documents often miss assumptions, edge cases, APIs, workflow states, and compliance details until development is already underway.