DAIVerse

AIVerse is a dynamic multimodal AI sandbox and creative platform that empowers users to experiment, create, and interact within intelligent virtual environments—melding visual storytelling, voice interaction, and avatar-driven immersion into one seamless experience.

Inspiration

In today’s AI landscape, creativity often stays within a single sensory channel. I envisioned a world where users can not just imagine—but live it. AIVerse was born from that urge: to build immersive, interactive AI spaces that are explorable, responsive, and uniquely yours.

What It Does

Visual Canvas Creation – Generate and explore AI-crafted 2D or 3D scenes.
Voice Interaction & Narration – Speak to elements and receive natural, AI-generated narration.
Avatar-AI Agents – Populate scenes with AI agents that respond, react, and engage.
Scenario Sharing – Create, share, and remix AI worlds within the community.

How It Was Built

Frontend: React.js, Three.js
AI Engines: Stable Diffusion (visuals), GPT-4 (behavior)
Speech Tech: Web Speech API, OpenAI TTS / ElevenLabs
Backend: Node.js / FastAPI
Storage: Supabase / Firebase

Challenges I Faced

Ensuring visual scene cohesion amid generative prompts
Delivering real-time voice interaction with low latency
Crafting a visually rich UI that remains usable for novices

Project Highlights

Fully functional AI-driven sandbox that blends visuals, speech, and AI agents
Real-time conversational avatars in custom scenes
Shareable user-generated "mini-worlds" accessible instantly

What ILearned

How to weave image generation, voice interactivity, and AI logic into a single flow
Prompt engineering for consistent scene generation
Balancing performance and experience in immersive applications