Yugma Platform
Yugma is an AI-native 3D studio — a professional web-based scene editor where natural language is the primary interface. Users describe what they want ("build a sci-fi hangar with neon lighting"), and an AI Director composes the scene using structured tool calls against a live Three.js scene graph.
The product is NOT a 3D editor with AI bolted on. It is an AI-native scene graph that happens to render in 3D.
Platform Overview
| Service | URL | Purpose |
|---|---|---|
| Studio App | yugmastudio-app.web.app | 3D scene editor with AI composer + real-time collaboration |
| Admin Panel | admin.yugma.studio | User management, provider config, usage stats |
| Documentation | docs.yugma.studio | This site |
| Landing Page | yugma.studio | Marketing site |
Tech Stack
| Layer | Technology |
|---|---|
| Frontend | React 18, TypeScript, Vite |
| 3D Engine | Three.js via React Three Fiber + Drei |
| State | Zustand (17 stores) |
| AI | Google Gemini, managed server-side (gemini-3.7-flash, env GEMINI_MODEL). It is the only inference provider; there is no user-facing model picker. |
| 3D Assets | First-party only: Library v2 (1,344-asset procedural world library, bundled), the yugma-lib v1 industrial catalog, and 20 NVIDIA Omniverse OpenUSD packs |
| Auth | Firebase Auth (Google OAuth only) |
| Database | Firestore + Realtime Database |
| Storage | Firebase Cloud Storage |
| Functions | Firebase Cloud Functions v2 (Node 22) |
| Hosting | Firebase multi-site hosting |
What Makes Yugma Different
-
Agentic AI loop — The AI calls tools, sees results (including generated IDs), and chains up to 5 dependent operations per turn (default 3). It builds coherent scenes, not random objects.
-
~60 tokens/object (measured) — The custom YSL serialization format fits a 50-object scene in ~3,000 tokens, letting the AI maintain full scene awareness.
-
Client-side spatial intelligence — Circle, grid, stack, spiral, and scatter patterns are computed with real trig, not hallucinated by the LLM.
-
16-phase roadmap — From scene editing (live) through digital twins, physics simulation, video reconstruction, and cinematic rendering (planned — earlier prototypes were removed in the 2026-07 cleanup).
Quick Links
- System Architecture — How everything fits together
- AI Pipeline — The 6-stage AI composition flow
- Phase Roadmap — All 16 phases with status
- AI Composer Guide — Using the AI chat interface
- Cloud Functions API — Backend reference
- Deployment — Build, deploy, and configure