Literature Assistant - Standalone

AI-powered scientific literature analysis assistant with comprehensive library management, community sharing, and multi-modal PDF parsing capabilities.

Features

Core Features

• PDF Upload & Parsing - Upload PDF papers with automatic text extraction
• Multi-Model AI Analysis - Support for 19+ LLM providers (Gemini, GPT, DeepSeek, Kimi, Claude, etc.)
• Real-time Streaming - Live streaming responses with heartbeat mechanism
• Multiple Analysis Modes - Deep dive, quick summary, multi-PDF comparison, image report, critique
• Custom Prompts - Create and manage your own analysis prompts

Library Management

• Folder System - Organize papers into custom folders (Want to read, Reading, Completed, etc.)
• File Preview - Quick preview panel with expandable view
• Grid/List View - Toggle between different view modes
• Search - Search across your library
• Batch Operations - Manage multiple documents

Community Sharing

• Public Sessions - Share your analysis with others
• Discover - Browse public literature analyses
• Copy/Fork - Copy public sessions to your library
• View/Copy Counts - Track engagement metrics

Advanced Features

• MinerU Integration - High-quality PDF parsing with figure extraction (图文解析)
• Figure-Aware Markdown - Render figures inline with AI responses
• URL Import - Direct import from arXiv, OpenReview, Nature
• Notes - Add personal notes to sessions
• Message Editing - Edit and resend messages
• Multi-format Download - Download as Markdown, PDF, or MinerU format
• Theme Toggle - Light/Dark mode support
• Drag & Drop - Drag files to upload

Authentication

• Google OAuth 2.0 - Secure authentication
• JWT Tokens - Stateless API authentication

Quick Start

Prerequisites

• Node.js 20+
• npm or yarn
• Google Cloud project (for OAuth)

Installation

cd literature-assistant-standalone
npm run setup    # Install dependencies for root, backend, and frontend

Configuration

Step 1: Environment Variables

cp .env.example .env

Edit .env with your configuration:

# ==================== Authentication ====================
GOOGLE_CLIENT_ID=your-google-client-id
GOOGLE_CLIENT_SECRET=your-google-client-secret
GOOGLE_CALLBACK_URL=http://localhost:3000/api/auth/google/callback

JWT_SECRET=your-random-jwt-secret-32-chars-minimum
SESSION_SECRET=your-random-session-secret-32-chars-minimum

FRONTEND_URL=http://localhost:5173

# ==================== LLM Providers ====================
# Option 1: JXTANG API Gateway (supports multiple models)
JXTANG_BASE_URL=https://your-api-gateway/v1/chat/completions
JXTANG_API_KEY=your-api-key

# Option 2: Direct Provider APIs
GEMINI_API_KEY=your-gemini-api-key
OPENAI_API_KEY=your-openai-api-key
DEEPSEEK_API_KEY=your-deepseek-api-key

# ==================== Optional Services ====================
# MinerU - PDF image extraction (图文解析)
MINERU_API_TOKEN=your-mineru-token

# DNS servers for MinerU CDN (optional)
DNS_SERVERS=8.8.8.8,1.1.1.1,114.114.114.114

Step 2: Configure Google OAuth

1. Go to Google Cloud Console
2. Create a new project or select existing
3. Navigate to APIs & Services → OAuth consent screen
   • User Type: External
   • Fill required fields (App name, emails)
4. Navigate to APIs & Services → Credentials
   • Click + Create Credentials → OAuth client ID
   • Application type: Web application
   • Authorized JavaScript origins: http://localhost:3000, http://localhost:5173
   • Authorized redirect URIs: http://localhost:3000/api/auth/google/callback
5. Copy Client ID and Client Secret to .env

Step 3: Configure LLM Providers

Edit backend/config/llm-providers.json to enable/disable providers:

{
  "defaultProvider": "gemini-3-pro-preview",
  "providers": {
    "gemini-3-pro-preview": {
      "name": "JXTANG",
      "baseUrlEnv": "JXTANG_BASE_URL",
      "apiKeyEnv": "JXTANG_API_KEY",
      "model": "gemini-3-pro-preview",
      "maxTokens": 32000,
      "multimodal": true,
      "enabled": true
    }
  }
}

Start Development Server

npm run dev

• Frontend: http://localhost:5173
• Backend: http://localhost:3000

Production Deployment

Using Docker

docker-compose up -d
docker-compose logs -f

Manual Deployment

npm run build
NODE_ENV=production npm start

Deploy to Zeabur

1. Push code to GitHub
2. Create project on Zeabur
3. Configure environment variables (change URLs to production domain)
4. Add production redirect URI to Google OAuth
5. Add Persistent Storage for /app/data

Architecture

literature-assistant-standalone/
├── frontend/                    # React + Vite + TypeScript
│   ├── src/
│   │   ├── pages/
│   │   │   └── LiteratureAssistant.tsx  # Main page (5600+ lines)
│   │   ├── components/
│   │   │   ├── ui/              # shadcn/ui components
│   │   │   ├── layout/          # Layout components
│   │   │   ├── literature-assistant/
│   │   │   │   └── FigureAwareMarkdown.tsx
│   │   │   └── MermaidChart.tsx
│   │   ├── stores/
│   │   │   └── literature-assistant-store.ts
│   │   ├── services/
│   │   │   ├── api.ts           # API client
│   │   │   └── auth.ts          # Auth service
│   │   └── hooks/
│   │       └── useTheme.ts      # Theme hook
│   └── ...
├── backend/                     # Express + SQLite
│   ├── api/
│   │   ├── auth-routes.js       # OAuth routes
│   │   ├── session-routes.js    # Session CRUD
│   │   ├── document-routes.js   # Document & figure routes
│   │   └── analysis-routes.js   # AI analysis, downloads, community
│   ├── lib/
│   │   ├── database.js          # SQLite wrapper
│   │   └── mineru-service.js    # MinerU integration
│   ├── middleware/
│   │   └── auth.js              # JWT middleware
│   └── config/
│       ├── llm-providers.json   # LLM configuration
│       └── prompts.json         # Analysis prompts
├── data/                        # Data directory
│   ├── database/                # SQLite database
│   ├── uploads/                 # PDF files
│   └── figures/                 # Extracted images
└── docker/                      # Docker configs

API Endpoints

Authentication

• GET /api/auth/google - Initiate OAuth
• GET /api/auth/google/callback - OAuth callback
• GET /api/auth/me - Get current user
• POST /api/auth/logout - Logout

Sessions

• GET /api/sessions - List sessions
• GET /api/sessions/:id - Get session with documents & messages
• POST /api/sessions - Create session
• PATCH /api/sessions/:id - Update session
• DELETE /api/sessions/:id - Delete session
• PATCH /api/sessions/:id/public - Toggle public status

Documents

• POST /api/documents - Upload document
• GET /api/documents/:id - Download document
• DELETE /api/documents/:id - Delete document
• POST /api/documents/:id/parse-images - Trigger MinerU parsing
• GET /api/documents/:id/figures - Get document figures
• GET /api/documents/:id/figures/by-label/:label - Get figure by label

Analysis

• POST /api/analyze - Stream AI analysis
• DELETE /api/messages/:id - Delete message
• POST /api/upload-from-url - Import from URL (arXiv, OpenReview, Nature)

Downloads

• GET /api/download/availability/:sessionId - Check download options
• GET /api/download/ai-response/:messageId - Download as Markdown
• GET /api/download/original-pdf/:documentId - Download original PDF
• GET /api/download/mineru-markdown/:documentId - Download MinerU Markdown
• GET /api/download/mineru-folder/:documentId - Download MinerU ZIP

Community

• GET /api/community/sessions - List public sessions
• POST /api/community/copy/:sessionId - Copy public session
• POST /api/community/view/:sessionId - Record view
• GET /api/community/settings - Get share settings
• POST /api/community/settings - Update share settings

Custom Prompts

• GET /api/custom-prompts - List custom prompts
• POST /api/custom-prompts - Create custom prompt
• DELETE /api/custom-prompts/:id - Delete custom prompt

Environment Variables

Variable	Description	Required
GOOGLE_CLIENT_ID	Google OAuth Client ID	Yes
GOOGLE_CLIENT_SECRET	Google OAuth Client Secret	Yes
GOOGLE_CALLBACK_URL	OAuth callback URL	Yes
JWT_SECRET	Secret for JWT tokens (32+ chars)	Yes
SESSION_SECRET	Secret for sessions (32+ chars)	Yes
FRONTEND_URL	Frontend URL for CORS	Yes
JXTANG_BASE_URL	API gateway base URL	At least one LLM
JXTANG_API_KEY	API gateway key	At least one LLM
GEMINI_API_KEY	Direct Gemini API key	Optional
OPENAI_API_KEY	Direct OpenAI API key	Optional
DEEPSEEK_API_KEY	Direct DeepSeek API key	Optional
MINERU_API_TOKEN	MinerU token for image parsing	Optional
DNS_SERVERS	DNS servers for MinerU CDN	Optional

Tech Stack

Frontend

• React 18 + TypeScript
• Vite
• Zustand (state management)
• Tailwind CSS + shadcn/ui
• Framer Motion (animations)
• React Markdown + KaTeX + Mermaid
• Recharts (analytics)

Backend

• Express 5
• SQLite3
• Passport.js (Google OAuth)
• JWT authentication
• pdf-parse (text extraction)
• MinerU API (image extraction)
• Sharp (image processing)

Troubleshooting

concurrently: not found

npm install   # In root directory

Cannot find module 'xxx'

npm run setup   # Reinstall all dependencies

Google OAuth redirect_uri_mismatch

• Verify redirect URI in Google Console matches exactly
• Check http vs https, correct port (3000 not 5173)

MinerU parsing fails

• Verify MINERU_API_TOKEN is set
• Check network connectivity to mineru.net
• Try configuring DNS_SERVERS if CDN is blocked

PDF upload fails

• Check file size (max 150MB)
• Ensure data/uploads directory is writable

License

MIT