I got tired of startup idea generators that spit out "Uber for dentists" and "AI-powered pet wellness dashboards." So I built the opposite, from a camper in Salmon, Idaho, on Zo Computer.

Launchling is a swipe-through deck of 500 hand-curated startup ideas. You open it, thumb through the cards, and the ones you match on open their full product requirements doc the second you swipe right. No email gate. No "unlock with email." No upsell. The dossier is right there — buyer, wedge, moat, validation path, pricing shape, build shape. The kind of doc you'd hand a contractor.

Every idea has a named buyer, a specific revenue model, and a reason it hasn't been built yet — because the interesting opportunities have friction, not frictionless market-fit theater. The corpus spans 20 categories (compliance tooling, rural logistics, creator infrastructure, artisan trades, and the other corners where real businesses live).

Creative vision. The deck is the product. The UI is built as a literal paper artifact — vellum, plum, brass, hairline rules, Fraunces + Inter Tight + JetBrains Mono — not an iOS sheet, not a SaaS template. Opening a PRP should feel like opening a physical dossier, not a modal.

Built entirely on Zo. One workspace holds the app, the 509 markdown PRPs, the Remotion-rendered trailer, the hosting config, and the custom domain binding. PRPs are plain .md files the server reads at request time — edit one, it's live on the next request. No CMS, no database, no Vercel, no deploy pipeline. zosite.json, a service, done. The 500-card corpus was generated, critiqued, and pruned using parallel Zo child-invocations — the same workspace that serves the app is the one that wrote its contents.

Try it: https://www.launchling.app

Swipe right on three cards, read the dossiers, and tell me if this looks like slop.

Built solo. Cursor, Claude, Zo. Lucky the border collie on the floor. Social posts: https://x.com/JeffKazzee/status/2046995008310382820, https://www.facebook.com/share/v/1ZS1b4DHNf/