Home matching for buyers and renters

Leo.casa is a property search app for people choosing a home in Spain, and the agents helping them. It brings a household’s routines, budget and priorities into the search: daylight, working from home, outdoor space, accessibility and neighborhood fit.

Property analysis is shared across searches. Changing a household’s priorities reranks the same analysis through deterministic scoring. The source includes the original intake, recommendations, property reports, maps, voice flows, data adapters and ingestion tools.

Original recommendation screen with fictional homes

The example uses fictional purchase homes and actual deterministic scores.

First result

Use Node 22.18 or later in the Node 22 line:

git clone https://github.com/latintzar/home-matching.git
cd home-matching
node --experimental-strip-types tools/first-match.mjs
node --experimental-strip-types --test tests/first-match.test.mjs

The original scorer ranks the courtyard home 84 for a buyer working from home. Choosing outdoor space and a pet makes the terrace home rank first at 90. A house-share fails the profile’s dealbreaker. A private receipt records the inputs, source and results.

Open the original recommendation scene:

cd starter
npm ci
npm run dev

Visit http://127.0.0.1:3000 and switch between the two profiles. The example runs locally with authored listings and analysis scores. The full application’s provider setup is described separately.

Build with it

Original code: MIT, copyright Peter Neyra. Third-party notices cover dependencies and fonts. Provenance records the source pin, authored examples and retained historical experiments.

made by atelier neyra

Search documentation

Type to find a guide or section.

↑ ↓ to choose · Enter to open · Esc to close