WonderLab
Voice-controlled VR world-building with AI
2020
Note: Built during the pandemic when everyone was losing touch with the places they loved—restaurants, parks, friends' houses. We wanted to let people recreate those spaces in VR, accessible to anyone regardless of technical skill. Just tell Rollanda what you want and she makes it appear.
WonderLab places users in a digital realm accompanied by Rollanda, an AI assistant robot who can make anything you desire appear in the space. She has access to a large library of assets—rooms, trees, furniture, paintings, kitchen items—and responds to natural voice commands. Say “I want a room” and she’ll make one appear with a cheerful “voila!”
The idea was to let anyone, at any skill level, reimagine their reality. Missing your family’s Thanksgiving table during quarantine? Recreate it without any coding or modeling experience. Want to turn your neighbor’s house into a dungeon? Go for it.
We integrated three IBM Watson services (Assistant, Speech-to-Text, Text-to-Speech) to create something novel—most Watson use cases involve data collection or automation, but we wanted to see what happens when you use it for creative expression. Users can also scale and rotate objects in real-time using the Oculus controllers, with distance between controllers determining scale.
Winner: IBM Community Response to COVID-19 at HackGT 7.
Built with Patrick Molen, Roshaan Siddiqui, and Ines Said.
key features
- Voice-controlled object spawning via IBM Watson Assistant
- Speech-to-text and text-to-speech for natural conversation with AI assistant "Rollanda"
- Real-time object scaling and rotation using Oculus Touch controllers
- Library of ~50 spawnable assets with color variants
technical details
Built in Unity for Oculus Quest. Integrated IBM Watson (Assistant, Speech-to-Text, Text-to-Speech). 3D assets created in Maya and textured in Substance Painter. Custom controller-based scaling system measuring real-time distance between controllers.