What it does

Augmented reality 3D room imaginer that takes in one or multiple images of a room, has Gemini 3 Pro Preview identify objects and generate a room with Three.js! The user can AI optimize their room just with words (ex. colour code everything, what if I could have more room near the window, etc.), search semantically for items to better fit their vision for their room, and even move in first-person views to navigate their dream room come to life!

Challenges we ran into

JSON errors from getting exact, strict strings from Gemini's API

What's next for Riko

Better 3D models for the objects and better object movements.

Built With

Share this project:

Updates