Project Description
I’m building the first piece of ShadowVerse, a social platform where people step into worlds instead of scrolling through feeds. The focus of this opening milestone is a powerhouse AI-driven avatar creator that lets users sculpt every detail of their digital self and then talk with it through a simple text chat.
Scope of work
The tool must give full control over appearance and behavior: facial structure, body proportions, skin options, hairstyles, outfits, accessories, plus the ability to assign idle and expressive animations. When a user tweaks a slider or selects an item, the change should appear instantly in a real-time 3D viewport (WebGL, Unity, Unreal Engine or comparable). Once the avatar is saved, a lightweight AI layer should enable basic conversational chat—no voice yet, just responsive text bubbles with appropriate idle gestures.
Technical needs
• Real-time 3D rendering suitable for web and mobile
• Clean JSON or GraphQL endpoint that stores the avatar’s parameter set and retrieves it on demand
• Export capability (GLB/FBX) so the same avatar can later enter our virtual rooms and mood-based worlds
• Documentation that explains how to extend the parameter list and plug in more advanced interaction models later
Acceptance criteria
1. A user can create an avatar from scratch, modify at least 30 distinct visual attributes, pick or upload textures, and preview looping animations.
2. Saved avatars load in under three seconds on a mid-range phone.
3. The embedded chat box answers user input within one second and triggers a context-appropriate idle or acknowledgment animation.
4. All code is handed over in a version-controlled repository with setup instructions that reproduce the demo with a single command.
This is the cornerstone feature that will power every other experience in ShadowVerse, so code quality and extensibility are critical. Let’s bring these avatars to life.