Discovery
Requirements, research, scope, and goals.
The phase where I decide what to build and why — before a single component exists. Cutting this short is the most expensive mistake I make, so I treat it as real work.
Overview
Discovery turns a vague idea into a bounded, justified scope. I gather requirements, research the problem and prior art, sketch the shape of the thing, and write down goals I can later check against. The output isn't code — it's clarity.
Goals
- Understand the real problem and who has it.
- Define scope: what's in, what's explicitly out.
- Set measurable goals so "done" is unambiguous.
- Surface constraints (technical, time, design) early, while they're cheap.
Activities
Tools & skills
- Excalidraw for quick flow and architecture sketches.
- Figma when discovery already needs a visual reference.
- An agent (Claude Code) to research prior art and pressure-test scope.
Deliverables
- A short requirements and scope note (in/out, non-goals).
- Rough flow or architecture sketches.
- Written goals and success criteria.
How Morpha UI did it
Morpha UI started as "I keep rebuilding the same stack every project — make it once." Discovery scoped that into a concrete shape: a Fumadocs documentation site that doubles as a personal design system, component library and knowledge base. I scaffolded the Fumadocs starter to get a feel for the content pipeline, then wrote a macro implementation plan that broke the work into ordered steps — each later step (tokens, components, CLI, content) traces back to a boundary drawn here.
Checklist
- The problem and audience are written down, not just in my head.
- Scope has an explicit boundary and a list of non-goals.
- Goals are measurable enough to judge "done."
- Major constraints and risks are surfaced.
- I have a rough sketch of the shape before moving to Design.