PRMPTR
Role: Design & Development
Client: Self
Turning image prompting into a fast, reusable workflow
PRMPTR is a web-based prompt builder that turns best-practice prompting into modular inputs - so creators can generate better image prompts faster, consistently.
PRMPTR started as a personal experiment - a small tool I built to make my own prompting workflows faster and more repeatable. After sharing it, the response was strong enough that I kept expanding it: refining the interface, formalizing prompt structure, and exploring how far a hyper-specific tool could go.
Built for my own workflow
I was spending too much time remembering prompt patterns that I knew worked. PRMPTR treats prompts like a design system - modular parts, clear ordering, and fast reuse - so I could focus on creative intent instead of syntax.
AI-Enhance
After selecting tags to build a structured prompt, PRMPTR includes an AI-Enhance step. This uses the OpenAI API to rewrite the assembled prompt into a more fluent, detailed version - preserving intent while improving clarity, specificity, and output quality.
A deliberate (and slightly silly) constraint
Almost every tag in PRMPTR includes a hover preview - and they all use the same subject: a goat. It’s playful, but intentional. Keeping the subject constant made it easier to confirm that prompt text was actually doing what it claimed, compare stylistic changes side-by-side without unnecessary noise, and iterate quickly while building and testing the system. The goats stuck around because they were fun.
My Entry-Point Into Coding
This project became my on-ramp into front-end development. I used modern tooling and AI-assisted coding to move quickly, learn by doing, and ship something real - embracing “vibecoding” while still caring deeply about structure, clarity, and UX.
More than anything, PRMPTR reflects a belief I’m excited about: a future where software doesn’t try to be everything - but instead does one thing extremely well for a very specific workflow.