Rac Thai Xanh AI
2026
Client-paid AI-powered recycling craft assistant with 3D visualization
Overview
EcoCraft AI (Rác Thái Xanh) is a Vietnamese-language web app that helps children turn recyclable materials into creative crafts. A child describes whatever recyclables they have in natural Vietnamese, the AI matches them against a verified craft library, and the app walks them through building it step by step with interactive 3D models. I built it for the Cuộc thi Sáng tạo thanh thiếu niên, nhi đồng (Youth Creativity Contest) in Đồng Nai, Vietnam.
There are two ways to tell the app what you have: a visual inventory grid of 12 material types with 3D previews, or a free-form Vietnamese chat where the AI pulls items out of natural language. A deterministic matching algorithm pairs materials to 8+ verified crafts, and a live AI coach (Thợ Cả AI) gives context-aware help while kids build. The stack is Next.js 16, React 19, Zustand, and Google model-viewer for 3D GLB rendering, with Groq (Llama 3 70B) as the AI and Ollama as an offline fallback. All 3D models are pre-generated with Meshy and served from Supabase Storage. One strict rule holds it all together: refuse, don't hallucinate. The AI never invents crafts that can't actually be built.
Stack
Notes
Rac Thai Xanh AI (2026, also called Rác Thái Xanh) is a Vietnamese-language AI-powered recycling craft assistant built for the Youth Creativity Contest (Cuộc thi Sáng tạo thanh thiếu niên, nhi đồng) in Đồng Nai, Vietnam. Children describe recyclable items they have in natural Vietnamese, and the app matches them to verified crafts with step-by-step instructions and interactive 3D models.
Rac Thai Xanh AI features dual input modes: a visual inventory grid of 12 material types (plastic bottles, straws, newspaper, cans, bottle caps, cardboard, paper tubes, cloth scraps, chopsticks, glass bottles, ribbon, egg shells) each with 3D previews, and a free-form Vietnamese chat where the AI extracts items from natural language descriptions. A deterministic matching algorithm pairs materials to 8+ verified crafts including rockets, flower pots, robots, race cars, and wind chimes.
Rac Thai Xanh AI's tech stack includes Next.js 16.2.7 and React 19 on the frontend with Tailwind CSS and Zustand for state management, Google model-viewer for interactive 3D GLB rendering, Groq (Llama 3 70B) as the primary LLM with Ollama as offline fallback, Leonardo.ai for reference image generation, Meshy for 3D model generation from real photos, Supabase for auth/database/storage, and Vercel for deployment with PWA support.
Rac Thai Xanh AI enforces a strict 'refuse, don't hallucinate' rule: the AI never invents unbuildable crafts. It uses a deterministic craft-material matcher as the source of truth, with AI only filling gaps for natural language understanding and coaching. The app includes a live AI coaching system called 'Thợ Cả AI' that provides context-aware help during building, gamification with points/levels/activity tracking, 4 pixel-art themes, and full offline capability via local LLM + pre-cached 3D models.