PaperDigest AI - Advanced Research Assistant

A professional, full-stack AI research platform designed to synthesize complex academic papers into actionable insights. PaperDigest AI leverages multi-staged LLM pipelines and RAG (Retrieval-Augmented Generation) to streamline literature reviews and technical deep-dives.

Features

📄 Intelligent Ingestion

• Structural Extraction - Automatically segments PDFs into logical sections (Abstract, Introduction, Methodology, etc.).
• Multi-pass Processing - Asynchronous extraction of titles, authors, and high-level paper metadata.
• Smart Formatting - Cleans raw PDF artifacts and multi-column layouts using PyMuPDF and Regex heuristics.

🧠 AI-Driven Analysis

• Hierarchical Summarization - Generates section-by-section summaries before synthesizing a final 'Global TL;DR'.
• Technical Deep-Dives - Specialized prompts for extracting technical methodologies, model architectures, and metrics.
• Metadata Tagging - Automatic identification of research datasets and software licenses used in the study.

🔍 Adv

anced Search & RAG

• Semantic Search - Search across your paper library using natural language, powered by vector embeddings.
• Side-by-Side Review - Compare AI summaries directly alongside the original PDF text.
• Context-Aware RAG - Injects the most relevant snippets from papers into LLM prompts for high-accuracy extraction.

📊 Review Dashboard

• Literature Review Matrix - A specialized view that aggregates summaries from all your papers into a comparison table.
• Researcher Notes - Add personal annotations alongside AI-generated insights.
• CSV Export - Export your entire literature review matrix for external analysis or reporting.

⚙️ Hybrid AI Backend

• Cloud Power - Native support for Google Gemini 1.5 Pro/Flash for high-accuracy processing.
• Local Privacy - Built-in integration with Ollama (Llama 3) for offline, privacy-first research analysis.
• Factory Pattern - Seamlessly toggle between LLM providers via environment configuration.

Getting Started

Prerequisites

• Docker & Docker Compose (Recommended)
• Git
• Google Gemini API Key (Optional, for cloud processing)
• Ollama (Optional, for local processing)

Installation

1. Clone the repository

   git clone https://github.com/Rishabds7/ai-research-assistant.git
   cd ai-research-assistant

2. Configure Environment Variables Create a .env file in the root directory (refer to .env.example in backend/):

   # Backend Settings
   GEMINI_API_KEY=your_api_key_here
   LLM_PROVIDER=gemini # or 'ollama'
   
   # Database Settings
   DB_PASSWORD=your_secure_password

3. Launch the platform

   docker-compose up --build

4. Access the application

   • Frontend: http://localhost:3000
   • Backend API: http://localhost:8000/api/

Technical Architecture

Core Technologies

• Next.js 15 & React 19 - Modern, responsive frontend with Server Components.
• Django & DRF - Robust Python backend for API management and data persistence.
• PostgreSQL & pgvector - High-performance vector database for semantic search.
• Celery & Redis - Distributed task queue for asynchronous AI processing.
• Tailwind CSS 4 - Premium UI styling with glassmorphism and modern aesthetics.

Key Components

AI Task Orchestration

PDF Upload → Celery Worker → PDFProcessor → LLM (Gemini/Ollama) → EmbeddingService → PostgreSQL

Vector Search Pipeline (RAG)

1. Embedding: Convert paper chunks into 384-dimensional vectors via all-MiniLM-L6-v2.
2. Storage: Vector data persists in PostgreSQL using the pgvector extension.
3. Similarity: Natural language queries are matched using Cosine Distance (<=>).

File Structure

research-assistant-mvp/
├── backend/            # Django Application
│   ├── core/           # Project settings & Celery config
│   ├── papers/         # Models, Views, and Tasks
│   └── services/       # Core AI/PDF Logic (LLM, Embeddings, PDFProcessor)
├── frontend/           # Next.js Application
│   ├── src/app/        # Pages & Layouts
│   ├── src/components/ # Atomic UI components
│   └── src/lib/        # API client & utility functions
├── docker-compose.yml  # Orchestration for DB, Redis, Worker, and Apps
└── README.md           # Professional Documentation

Browser Compatibility

• ✅ Chrome 110+
• ✅ Firefox 100+
• ✅ Safari 16+
• ✅ Edge 110+

Development & Testing

Key Backend Services

• services.pdf_processor.PDFProcessor - Text segmentation logic.
• services.llm_service.LLMService - Multi-provider (Gemini/Ollama) factory.
• services.embedding_service.EmbeddingService - Vectorization and RAG logic.

API Documentation

The API documentation is available via Swagger/ReDoc at /api/docs/ when the backend is running.

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-ai-tool)
3. Commit your changes (git commit -m 'Add amazing AI feature')
4. Push to the branch (git push origin feature/amazing-ai-tool)
5. Open a Pull Request

---

PaperDigest AI - Synthesizing deep research into actionable insights.