Oasis
Build custom 3D app interfaces by describing your vision to a real-time, voice-synced AI world-building agent.
YouTube Video
Project Description
Oasis turns any educational intake form into a living 3D world.
A teacher creates a Google Form. Oasis turns it into an interactive environment and generates an interactive environment, populated with activities, objects, and media, all built by agents in real time. Students don’t navigate a dashboard. They enter a world.
This is not a chat interface with a prettier skin. Oasis is generative UI as a primary operational layer: the interface itself is the output. Users create inside it using text, images, or voice, and the world responds. Text-to-3D, image-to-3D, voice-to-3D flows are live and composable.
The core technical bet is A2UI, agents that write functional UI, not just content. Oasis is the first application proving this pattern works at interactive scale.
What’s already built:
-Generative world creation from structured input (Google Form → 3D environment)
-Text, image, and voice creation flows inside worlds
-Rich media support: images, video, sound embedded in-world
-Showcase worlds and PortalZero / Conjure for spawning interactive objects
Roadmap: auth/login, credits for compute-intensive generation, BYOK/OAuth for enterprise access.
Live demo:
04515.xyz
Prior Work
during the hackathon we created the Gemini Live agent, the Google Form to 3D interactive world feature, using A2UI.