RACEprompt
The prompt optimization layer between you and your AI. Role · Action · Context · Expectation — four fields, 75+ templates, 11 languages, Smart Clarify. One subscription covers every platform.
The ambitious test. Monolith's skeleton kept intact — six 100vh slabs, one idea per viewport, 16vw hero, horizontal apps carousel, scroll-snap rhythm. Spatial's vocabulary bolted on only at handshake moments — places where the marketing site would otherwise hand off to the product. The rule that makes it work: if every slab goes glass, it goes mushy. So atmosphere is page-wide, but glass only shows up where the site is selling a surface the user will eventually touch — the apps, the CTA.
0.001%: unity of place > theatrical contrast.0.001%: emphasis requires restraint; if everything sparkles, nothing does.mask-composite: exclude — reads as emission from inside the panel, not a stroke applied on top. The effect is the visual argument that this is the stuff you touch. 0.001%: craft is in the 1px.0.001%: motion with purpose, stillness with meaning.0.001%: preserve the signature, translate the context.A field-tested approach to AI at work — from someone who ships three apps while running a PMO. No decks between your brief and a working prototype.
Mine are shipping with it. If your organization is still in the "let's schedule a workshop" phase, I can save you a quarter by sending you back a working prototype tomorrow instead.
The prompt optimization layer between you and your AI. Role · Action · Context · Expectation — four fields, 75+ templates, 11 languages, Smart Clarify. One subscription covers every platform.
Twelve breathwork patterns. Zero accounts. Zero tracking SDKs. First-class Apple Watch — not a remote control. Built for people who actually have jobs.
Focus. Fight. Finish. Identity-based accountability, built with Jada Sullivan. iOS + Android + a Watch complication that reads clean at 41mm.
Who is the model playing? Name it precisely. "A senior technical writer" beats "an expert".
What is the verb? "Draft", "compare", "refactor", "debunk". Precision over politeness.
What does it need to know that isn't in the prompt? Constraints, artifacts, audience.
What does "finished" look like? Length, format, tone, what to omit. This is where most prompts die.
Every other consultant wants a discovery call. I want your problem in writing and twenty-four hours. Most clients use what I ship the same day to get their team aligned on what they actually want to build.
Not a mockup. Not a concept deck. Real interactive code with your data.
Plain-English rationale — what was built and the reasoning behind it.
Anyone on your team can use it the same day — no onboarding call.
Iteration priced by outcome, not by the hour. Ship, scale, or walk away clean.
Pick your tool and go. The tools are freaking incredible right now. I'm a few steps ahead — not above — and the things I ship are the notes I'd give you anyway.