We build in AI WYSIWYG
A browser studio that takes a non-technical owner from no brand kit to a published five-page site in about forty minutes.
The problem
A small-business owner with no logo, no palette, and no spec has two bad options. Page builders hand over empty text and image boxes with no help on type, color, spacing, or photos. Whole-page AI generators burn tokens and overwrite every edit. Either way she ends up waiting on a developer.
What we built
AI WYSIWYG asks two questions, takes two or three sites she likes, and has her tap four either-or choices. From that it produces three distinct brand systems, proposes a five-page structure instantly, and drafts every page in her voice with every photo slot already filled. She edits by tapping or asking, never dragging, and nothing she does can go off-brand.
- Three brand systems generated from reference sites and four taps
- Five-page structure proposed by rules, with blocks pre-chosen
- Copy drafted in the brand voice, banned-phrase check built in
- Every photo slot filled from stock with the brand tint applied
- One-tap page check with plain-language fixes
- Lead form reaches her inbox and a Google Sheet in seconds
Project context
Stack: TypeScript, Firebase magic-link sign-in, Firestore, Claude and Gemini for brands and copy, Playwright for reference-site extraction, Unsplash for photos, static publish to a custom domain. Target: a publishable five-page site in about forty minutes for under five dollars of API cost. Status: invite-only; the first client site, a medical concierge practice, was built on it.
Let’s look at your software
Show us your workflow and what you would like to change.
Book a free 30-minute call