Racep Labs

Märchenfuchs tells children bedtime stories that exist only once: with their own name, their favorite topics, matched to their age. Parents set up once what their child loves; the app turns that into a new story every evening.

There is more behind it than a text generator: stories follow age-appropriate narrative patterns, and parents stay in control of topics and length. So quality does not come down to luck, a dedicated evaluation stage scores every story against a fixed rubric, with a set of reference cases as the benchmark.

The app is built and running, the payment model is in place, and a separate development instance takes new features before they go live. Right now we are looking for the first beta families before opening up.

We are building Märchenfuchs as a team of three under Racep Labs. I own the technical side: the Flutter app, the Supabase backend and the story pipeline behind it.

Lifemate bundles the areas otherwise scattered across five apps (calendar, fitness and nutrition tracking, finances and study notes) in one interface, with an AI layer on top. Instead of only collecting data, Lifemate helps with planning: training and nutrition plans are generated and adjusted to your own goals.

The core by now is the daily-routine engine: a scheduler lays out the day, streaks keep habits going, and a persistent memory means the app knows across sessions what you are currently working on. As a PWA it works offline too; since the last redesign the interface follows a monochrome token system instead of assembled one-off colors.

Technically Lifemate is a monorepo with two frontends (a web app for desktop and mobile, a touch variant for tablets and smart displays) on a shared codebase: Next.js and React on the front end, Express and Firestore on the back end, plus end-to-end tests with Playwright.

Lifemate is my own project at Racep Labs, from architecture to deployment.