All projects
02 / Product engineering · AI workflowsLive product

Project Brainstormer

From an interesting idea to a project worth testing.

A workspace to discover ideas, develop structured briefs, compare alternatives and record validation experiments. AI proposes changes; the person decides what to keep.

Open live product
VueJavaScriptSupabaseAI APIsVercel
Project BrainstormerPB ↗
01 / DISCOVERA question
worth asking.
02 / SHAPEMake it
tangible.
03 / VALIDATETest the
assumption.

Product illustration · Explore the actual interface through the live product link.

01 / Context

Start with the problem.

Generating another idea is easy. Understanding the problem, comparing options and deciding what to test next is harder. Ideas need structure and a path to evidence.

My contribution

Personal product: rebuilt an earlier commerce project into a focused idea-development workspace, including discovery, private project libraries, revisions and deployment.

  • Keep private projects isolated between users.
  • Make AI suggestions reviewable before applying them.
  • Preserve manual work when AI is unavailable.
  • Bound generation cost and request duration.
02 / System overview

Follow the flow.

A deliberately simplified view of the product. The decisions below explain the boundaries that matter.

03 / Engineering judgment

The decisions behind the interface.

01

Suggestions are proposals

AI edits do not silently overwrite a project. The user reviews a proposal, preserving ownership of the brief and the reasoning behind it.

02

Useful without generation

Manual editing, uploads and Markdown export remain available independently of AI. The core workspace retains value when a provider fails or an allowance is reached.

03

Validation before conviction

Structured briefs, alternatives and experiments treat concepts as hypotheses. A polished AI answer is not presented as proof of market demand.

04 / Evidence & boundaries

What exists. What remains open.

  • A deployed product with illustrated concepts and a private project library.
  • The source includes checks for isolation, revisions, conflicts and generation quotas.
  • Production is connected to the GitHub main branch through Vercel.

Known limitations

  • Concepts and AI output are hypotheses, not verified market research.
  • No public conversion, adoption or business-outcome measurements are claimed.
  • Source is currently private; the live product and this architecture overview are available for review.
05 / The next useful step

Keep asking better questions.

Measure whether a structured brief helps users design a better first validation experiment.

Next project / Planned

Evidence Desk