Product Case Study

Flake —
an AI-powered real estate marketplace

A real estate marketplace for Kosovo built to let people search for a home in plain language instead of fighting with filters.

What it is

Flake is a real estate marketplace for the Kosovo market. Instead of the usual grid of filters, Flake lets people search the way they'd actually describe what they want — something like "shtëpi 3 dhoma pranë qendrës nën €150 mijë" — and returns matching listings from natural language.

What we built it with

Next.js, React, and TypeScript on the frontend with Tailwind CSS, Supabase (Postgres) for data and image storage, and Mapbox GL for map-based browsing. The AI-powered natural-language search is the core feature — parsing a plain-language query into structured filters (rooms, location, price) behind the scenes. Deployed on Vercel.

Next.js
React, TypeScript & Tailwind CSS
Supabase
Postgres backend & image storage
AI search
Natural-language query matching

Flake is a WDM-built platform, not a third-party client engagement — included here as an example of what we build when we own the product end to end.

Want something built to this standard?

Get a free digital audit and see exactly where your current site is losing you clients.

Get your free audit →

View the live site →