Three.ws
Generate, animate, and embed interactive 3D avatars powered by LLM intelligence and web-native rendering directly from text or images.
Three.ws is an open-source, browser-native platform designed to give AI agents an interactive 3D embodiment. By transforming text prompts, sketches, or photos into textured, rigged, and animated GLB models via its Forge generator, it enables developers to pair visual digital humans with LLM reasoning, memory, on-chain wallets, and remote MCP connectivity. Built on Three.js and deployable anywhere via lightweight web components, Three.ws eliminates plugin dependencies and heavy server overheads to deliver seamless, real-time 3D AI agent interactions across the modern web.
Características Principales
01254 GitHub stars
02Generates textured, rigged, and animated 3D GLB models from text, images, or sketches
03Integrates LLM reasoning, conversational memory, and on-chain wallet functionality into avatars
04Renders natively in web browsers using Three.js without requiring plugins or complex installs
05Embeds interactive 3D models into websites and markdown documents with a single web component
06Provides remote Model Context Protocol (MCP) server support and x402 payment integration
Casos de Uso
01Deploying interactive 3D digital human assistants and customer support agents across web applications
02Embedding lightweight, interactive 3D model previews into GitHub READMEs, documentation, and issues
03Powering autonomous on-chain AI agents with visual embodiments capable of managing transactions and user interactions