Vaisu - Text to Visual Intelligence

Transform text documents into interactive visual representations using AI-powered analysis.

Features

• Document Upload: Support for .txt, .pdf, and .md files
• Text Input: Paste text directly for instant analysis
• AI Analysis: Powered by Claude 3.7 and GPT-4.5 via OpenRouter
• Multiple Visualizations:
  • Structured View (default) - Document outline with summaries
  • Mind Map - Hierarchical concept visualization
  • Flowchart - Process flow diagrams
  • Knowledge Graph - Entity relationships
  • Executive Dashboard - KPIs and metrics
  • Timeline - Chronological events
  • And more...
• TLDR & Executive Summary: Instant insights
• Interactive Exploration: Zoom, pan, search, and filter
• Export: PDF, PNG, SVG, PPTX, HTML formats

Tech Stack

Frontend

• React 18 + TypeScript
• Vite
• Zustand (state management)
• TailwindCSS
• D3.js, Cytoscape.js, React Flow (visualizations)

Backend

• Node.js + Express
• TypeScript
• OpenRouter API (LLM access)
• PDF/DOCX parsing

Prerequisites

• Node.js 18+ and npm
• OpenRouter API key (get one here)

Quick Start

1. Clone and Install

cd vaisu
npm install

2. Configure Backend

Create backend/.env:

PORT=7001
OPENROUTER_API_KEY=your_key_here
OPENROUTER_BASE_URL=https://openrouter.ai/api/v1
APP_URL=http://localhost:7002

3. Start Development Servers

# Start both frontend and backend
npm run dev

# Or start separately:
npm run dev:backend  # Backend on http://localhost:7001
npm run dev:frontend # Frontend on http://localhost:7002

4. Open Application

Navigate to http://localhost:7002

Usage

1. Upload a Document or Paste Text
2. Wait for AI analysis (10-30 seconds depending on document size)
3. View TLDR and Executive Summary
4. Explore different Visualizations
5. Export your favorite views

Project Structure

vaisu/
├── frontend/          # React frontend
│   ├── src/
│   │   ├── components/
│   │   │   ├── upload/
│   │   │   ├── summary/
│   │   │   └── visualizations/
│   │   ├── stores/
│   │   ├── services/
│   │   └── App.tsx
│   └── package.json
├── backend/           # Express backend
│   ├── src/
│   │   ├── routes/
│   │   ├── services/
│   │   │   ├── llm/
│   │   │   ├── analysis/
│   │   │   └── visualization/
│   │   └── server.ts
│   └── package.json
├── shared/            # Shared TypeScript types
│   └── src/types.ts
└── package.json       # Root workspace config

API Endpoints

• POST /api/documents/upload - Upload document file
• POST /api/documents/analyze - Analyze document or text
• GET /api/documents/:id - Get document details
• POST /api/documents/:id/visualizations/:type - Generate visualization
• GET /api/health - Health check

LLM Models Used

Via OpenRouter:

• Claude 3.7 Haiku: Fast summaries, section analysis
• Claude 3.7 Sonnet: Executive summaries, relationships, recommendations
• GPT-4.5 Mini: Entity extraction, KPIs, signals
• GPT-4.5: Technical diagrams, complex analysis

Development

Build for Production

npm run build

Run Tests

npm test

Environment Variables

Backend (.env):

• PORT: Server port (default: 7001)
• OPENROUTER_API_KEY: Your OpenRouter API key (required)
• OPENROUTER_BASE_URL: OpenRouter API URL
• APP_URL: Frontend URL for CORS
• NODE_ENV: development | production

Troubleshooting

"OPENROUTER_API_KEY is required"

• Make sure you've created backend/.env with your API key

"Failed to upload document"

• Check file size (max 10MB)
• Ensure file type is .txt, .pdf, or .docx

Backend not starting

• Ensure port 7001 is available
• Verify env vars are loaded
• Check DynamoDB connection
• Verify backend is running on port 7001
• Check Node.js version (18+ required)

Frontend not connecting to backend

• Verify backend is running on port 3001
• Check Vite proxy configuration in frontend/vite.config.ts

Roadmap

• Real-time collaboration
• More visualization types (UML, Gantt, etc.)
• Advanced mind map interactions
• Custom themes and branding
• User accounts and document storage
• Mobile app
• Plugin system

License

MIT

Contributing

Contributions welcome! Please open an issue or PR.

Support

For issues or questions, please open a GitHub issue.