hommi

The problem

Finding a WG or flatshare in Germany is a minefield: fake listings, ghost landlords, mismatched roommates, and Vorkasse scams where you're asked to wire a deposit before you've ever seen the room. The signal-to-noise ratio is brutal, and the people most exposed — newcomers and students — are the least equipped to spot the fraud. hommi rebuilds the search around trust and fit instead of luck.

The approach

1. 01

   Swipe-based matching

   Rooms and roommates surface as a swipe feed, so browsing feels fast and low-friction instead of trawling endless list pages. Interest is mutual before anyone shares contact details.

2. 02

   AI recommendation engine

   A recommendation engine sorts each user's feed by fit — lifestyle, budget, location, and stated preferences — so the most relevant matches rise to the top instead of whatever was posted most recently.

3. 03

   KYC verification for trust

   Identity (KYC) verification gates the platform, so profiles are backed by real, checked people. That single layer removes most of the fake-listing and Vorkasse fraud that plagues the incumbents.

4. 04

   Location-aware & native, shipped solo

   PostGIS powers geospatial search — distance, neighborhood, and commute-aware results — and Capacitor ships the same codebase as native iOS and Android apps. I built the whole thing solo: schema, tRPC API, and app.