Advanced RAG System
For Your Documents
Upload documents, images, and scanned PDFs with OCR support. Organize in folders, search with context memory, and get AI-powered answers instantly. Hybrid BM25 + semantic search with automatic indexing. All processing stays on your machine.
Powerful Features
Everything you need for intelligent document Q&A, built with privacy and performance in mind
Privacy First
100% local processing. Your documents and queries never leave your machine. No cloud, no tracking, no data collection.
Hybrid Search (NEW!)
Advanced RAG with BM25 keyword search + semantic embeddings. Reciprocal Rank Fusion combines the best of both worlds.
Folder Organization (NEW!)
Organize documents into custom folders with colors. Human-readable paths like docs/sales/ instead of cryptic IDs. Visual folder management interface.
Folder-Specific Search (NEW!)
Search only within selected folders for 70% faster results. Targeted searches give more relevant answers from specific document categories.
Auto-Indexing (NEW!)
Documents are automatically indexed after upload. No manual commands needed! Folder-specific indexes for faster search performance.
Conversation Memory (NEW!)
AI remembers your chat context! Ask follow-up questions naturally. Last 6 messages preserved for contextual understanding.
Chat History Sidebar (NEW!)
Visual sidebar showing all session questions with timestamps. Click to jump to previous answers. Clear history to start fresh topics.
Simplified Q&A Interface (NEW!)
Clean, focused question input. Removed unnecessary preview panel. Auto-resizing textarea with keyboard shortcuts for faster interactions.
OCR Support (NEW!)
Upload images (PNG, JPG, TIFF) and scanned PDFs! Tesseract OCR automatically extracts text from images with confidence scoring and preprocessing.
Authentication System (NEW!)
Secure JWT-based authentication with NextAuth.js. Protected routes, role-based access, and session management. Beautiful custom sign-in UI.
Auto Re-indexing
Active — Automatically re-index documents when files change in the docs folder. Keeps search results fresh without manual action.
Encrypted Vault
Active — AES-256 encrypted storage for sensitive documents. Keys remain local to your machine for maximum privacy.
Voice Q&A
Active — Hands-free question answering using speech recognition and local models for quick, conversational access.
Analytics Dashboard
Active — Track searches, document usage, and identify knowledge gaps to understand what your users are asking.
Smart Chunking
LangChain RecursiveTextSplitter with 1000-char chunks and 200-char overlap for optimal context preservation.
Intelligent Reranking
RRF algorithm fuses keyword and semantic results, then optimizes context window with top-5 chunks.
Source Confidence Scoring
Every answer includes source attribution with confidence scores. Know exactly where information comes from.
How It Works
Get started in minutes with our simple 3-step process
Upload Documents
Drag and drop your PDFs, Word docs, Excel files, images, or any supported format. OCR automatically extracts text from images and scanned PDFs. Our system indexes everything instantly.
Ask Questions
Type your questions in natural language. Our AI understands context and finds the most relevant information from your documents.
Get Instant Answers
Receive accurate, source-attributed answers in seconds. Export, share, or refine your queries as needed.
System Architecture
Advanced RAG pipeline with OCR, hybrid search, intelligent reranking, and automatic indexing
Who Is This For?
Perfect for anyone who works with documents and needs quick, intelligent answers
Researchers
Quickly find information across research papers, journals, and academic documents.
- Literature reviews
- Citation finding
- Data analysis
Students
Study smarter by querying textbooks, lecture notes, and study materials.
- Exam prep
- Assignment help
- Quick lookups
Businesses
Access company knowledge bases, policies, and documentation instantly.
- HR policies
- Product docs
- Training materials
Developers
Navigate technical documentation, API references, and code repositories.
- API docs
- Code examples
- Best practices
Simple, Transparent Pricing
Free and open source. Run it anywhere, anytime.
Free
Perfect for personal use
- Unlimited documents
- All file formats
- 100% local processing
- Community support
- Open source
- Priority support
- Custom integrations
Pro
For teams and businesses
- Everything in Free
- Priority support
- Custom integrations
- Advanced analytics
- Team collaboration
- SSO & SAML
- SLA guarantee
Enterprise
For large organizations
- Everything in Pro
- Dedicated support
- On-premise deployment
- Custom AI models
- Training & onboarding
- Custom SLA
- White-label option
Honest Limitations
We believe in transparency. Here's what you should know before getting started.
Requires Local Hardware
You need a computer with sufficient RAM and CPU to run Ollama and the AI model. Minimum 8GB RAM recommended.
OCR Processing Time
Images and scanned PDFs are processed with OCR, which may take longer than regular text extraction. Processing time depends on image quality and size.
File Size Limits
Individual files are limited to 10MB for optimal performance. Larger files may need to be split.
No Cloud Sync
Since everything is local, there's no automatic sync across devices. You manage your own data.
Get In Touch
Have questions? Want to contribute? We'd love to hear from you.
Contact Information
About the Author
Built with ❤️ by developers who believe in privacy-first AI solutions. We're passionate about making powerful AI tools accessible to everyone without compromising data security.
This project is open source and welcomes contributions from the community. Join us in building the future of private, local AI.
Stay Updated
Subscribe to our newsletter for updates, tips, and new features.
We respect your privacy. Unsubscribe at any time.