Generate consistent, professional 3D cartoon illustrations for yoga scripts. Paste a yoga script, customize the character, and generate start/end frame images for each pose using AI.

• Script Parser — Paste any yoga script; auto-detects intro, poses (numbered), and outro sections
• Prompt Generator — 6-layer prompt architecture with a 15-pose knowledge base for accurate yoga poses
• Multi-Provider Image Generation — Google AI (Gemini), Vertex Key, and Gommo AI
• Character Consistency — Reference images + detailed character description for consistent look across all frames
• Sequential Queue — Rate-limited, retry-capable queue with cancel support
• Persistent State — Hybrid localStorage + IndexedDB; survives page refreshes
• ZIP Export/Import — Download project as ZIP or load a previous project from ZIP
• Veo 3 Video Prompts — Auto-generated video transition prompts for each pose

# Install dependencies
npm install

# Start development server
npm run dev

# Build for production
npm run build

The app supports three providers. Enter your API key in the Settings panel:

API keys are stored in your browser's localStorage only — never sent to our servers.

src/
├── main.js                  # App entry point + UI wiring
├── modules/
│   ├── imageGenerator.js    # Multi-provider image generation
│   ├── imageQueue.js        # Sequential queue with retry
│   ├── logger.js            # Production-safe logging
│   ├── promptGenerator.js   # 6-layer prompt builder
│   ├── scriptParser.js      # Yoga script parser
│   └── state.js             # Persistent state (localStorage + IndexedDB)
├── styles/
│   └── index.css            # All styles
api/                         # Vercel serverless proxy functions
├── google-ai.js
├── vertex-key.js
└── gommo.js

Configured for Vercel. Push to GitHub and connect via Vercel dashboard.

• Proxy functions in /api/ handle CORS for external API calls
• vercel.json configures rewrites and build settings

MIT