PropertyLens

Inspiration

Home listings show bedrooms, square footage, and photos, but they don’t always answer the questions that shape daily life: How long is the commute? What’s nearby? Will my furniture fit? We built PropertyLens to bring more of that context into one home search.

What it does

PropertyLens combines listings from Zillow and Redfin, lets users search in plain English, and brings together property details and neighborhood insights such as schools, commute estimates, groceries, crime information, and natural hazards. Users can save homes and try furniture layouts in an interactive 2D and 3D room planner.

How we built it

We built the app with React and Vite, an Express backend, and Three.js for the room planner. We used Apify actors to gather home listing data from Zillow and Redfin, then combined and organized the results. We used CubiCasa to detect rooms and features in uploaded floor plans and convert them into 3D layouts. The planner also supports estimated layouts and a searchable IKEA furniture catalog.

Challenges we ran into

Turning CubiCasa floor plans into usable room layouts was difficult. It didn’t always recognize doors, windows, or room boundaries correctly, and those tracing errors made it hard to convert pixels into reliable real-world measurements. Getting the feet-to-pixel scale right was especially challenging when the original plan wasn’t traced cleanly. We also had trouble finding complete, reliable data for some properties.

Accomplishments that we're proud of

We brought home search, property context, saved favorites, and hands-on space planning into one experience. Users can go from finding a listing to exploring how a home might work for them.

What we learned

We learned that different AI models work best with different input formats and data types. We had to adapt and validate the information we sent so each model could handle it reliably. We also learned to make clear where property information comes from and when it’s an estimate.

What's next for PropertyLens

We’d like to improve floor plan accuracy, make the 3D render more realistic, expand listing and neighborhood data coverage, and make it easier to compare homes based on each buyer’s priorities.

Apify Actor IDS

Neighborhood Crime Safety: QMw7eHkWvgqxhnd0C IKEA Furniture Finder: 1s1g2xDQFnPD0zt6C

Built With

• apify
• express.js
• fema
• google-maps
• greatschools
• javascript
• node.js
• openai-api
• react
• react-three-fiber
• redfin
• rentcast
• three.js
• us-census-geocoder
• vite
• zillow