Marginalia

A reading assistant that turns documents into conversational agents. Upload PDFs or texts, and each source becomes an AI agent that responds from its own perspective when you select passages.

How It Works

1. Upload documents - Add PDFs or paste text as sources
2. Identity extraction - The system extracts each source's core commitments, vocabulary, and argumentative style
3. Read and select - Browse your documents and select passages of interest
4. Agent responses - Relevant sources respond from their own perspectives, not as neutral summaries

The interface implements a "moderated seminar" where your sources can participate in discussion while you read or write.

Requirements

• Node.js 18+
• Google Gemini API key

Setup

# Install dependencies
npm install

# Create environment file
cp .env.example .env.local

# Add your Gemini API key to .env.local
# GEMINI_API_KEY=your_key_here

# Run development server
npm run dev

Open http://localhost:3000

Project Structure

src/
├── app/                    # Next.js pages and API routes
│   ├── page.tsx           # Main reading interface
│   ├── library/           # Document library
│   ├── sources/           # Source management
│   ├── workspaces/        # Workspace switcher
│   └── api/               # Backend endpoints
│       ├── chat/          # Agent chat
│       ├── parse-pdf/     # PDF text extraction
│       ├── extract-identity/  # Source identity extraction
│       └── prefilter/     # Relevance filtering
├── components/            # React components
├── hooks/                 # Custom React hooks
└── lib/                   # Utilities and business logic
    ├── gemini.ts          # Gemini API client
    ├── types.ts           # TypeScript interfaces
    └── prompts.ts         # LLM prompt templates

Tech Stack

• Next.js 14 (App Router)
• TypeScript
• Google Gemini AI
• pdfjs-dist for PDF parsing
• Tailwind CSS
• LocalStorage for data persistence

License

MIT