ragadoc

An AI document assistant that answers questions about your PDFs with citations and highlights them directly in the document.

Ragadoc is a privacy-first Streamlit application that lets you chat with your documents using locally-run AI models. Ask questions, get grounded answers with citations, and see exactly where the information comes from with automatic PDF highlighting.

✨ Key Features

• 🤖 AI Document Q&A - Ask natural language questions about your PDFs
• 📍 Citation Grounding - Every answer includes specific citations from your document
• 🎯 PDF Highlighting - Citations are automatically highlighted in the original PDF
• 🔒 Complete Privacy - Uses only local AI models, your documents never leave your computer
• ⚡ Fast Processing - Optimized document parsing and retrieval system
• 🌐 Easy Web Interface - Simple Streamlit app, no technical knowledge required

Main chat interface with document upload and conversation

Document analysis with citations and highlighted responses

⚠️ Warning: Proof of Concept, Early Development

This application is currently in early development and should be considered a proof of concept. Features may be incomplete, unstable, or subject to significant changes. Use at your own discretion and expect potential bugs or breaking changes in future updates.

🚀 Quick Start

Model Selection Guide

Choose models based on your system capabilities:

Model Type	Model Name	Size	RAM Required	Use Case
Embedding	nomic-embed-text	~274MB	1GB	Recommended - General purpose
Embedding	all-minilm	~23MB	512MB	Lightweight alternative
Chat	qwen3:14b	~8.5GB	16GB	Recommended
Chat	llama3.1:8b	~4.7GB	8GB	Balanced option
Chat	mistral:latest	~4.1GB	8GB	Quick responses
Chat	phi3:mini	~2.3GB	4GB	Low-resource systems

Prerequisites (Required for Both Installation Methods)

1. Install Ollama (for local AI models):

# macOS
brew install ollama

# Or download from https://ollama.com

2. Start Ollama and install required models:

ollama serve

# Install embedding model (required)
ollama pull nomic-embed-text

# Install a chat model (see recommendations above)
ollama pull qwen3:14b

Installation Options

Choose your preferred installation method:

Option 1: Direct Installation

Additional Prerequisites:

• Python 3.8+

Installation Steps:

1. Clone the repository:

   git clone https://github.com/yourusername/ragadoc.git
   cd ragadoc

2. Install Python dependencies:

   pip install -r requirements.txt

   Or with conda:

   conda env create -f environment.yml
   conda activate ragadoc

3. Launch the application:

   streamlit run app.py

4. Open your browser to http://localhost:8501

Option 2: Docker Installation

Additional Prerequisites:

• Docker and Docker Compose

Installation Steps:

1. Clone the repository:

   git clone https://github.com/yourusername/ragadoc.git
   cd ragadoc

2. Start with Docker Compose:

   docker-compose up

3. Open your browser to http://localhost:8501

📖 How to Use

1. Upload a PDF - Drag and drop or browse for your document
2. Select AI Model - Choose from your locally installed Ollama models
3. Start Chatting - Ask questions about your document in natural language
4. View Citations - See highlighted text in the PDF that supports each answer
5. Explore - Continue the conversation to dive deeper into your document

🏗️ Architecture

Ragadoc uses a modern RAG (Retrieval-Augmented Generation) architecture:

PDF Upload → Text Extraction → Chunking → Vector Embeddings
                                               ↓
User Question → Semantic Search → Context Retrieval → AI Response
                                               ↓
                                    Citation Highlighting

Tech Stack:

• Frontend: Streamlit web interface
• AI Models: Ollama (local LLMs)
• Vector DB: ChromaDB for semantic search
• PDF Processing: PyMuPDF4LLM for structure-aware extraction
• Embeddings: nomic-embed-text model

🐛 Troubleshooting

Common Issues

Ollama Connection Error

# Verify Ollama is running
curl http://localhost:11434/api/version

# If using Docker, ensure external access
OLLAMA_HOST=0.0.0.0:11434 ollama serve

Slow Performance

• Try next smaller model
• Reduce chunk size in expert RAG settings
• Ensure sufficient RAM is available

📄 License

This project is licensed under the GPL License - see the LICENSE file for details.

🙏 Acknowledgments

• Ollama for making local AI accessible
• Streamlit for the amazing web framework
• PyMuPDF for PDF processing
• ChromaDB for vector storage

---

⭐ Star this repo if Ragadoc helps you work with your documents more effectively!