🤖 A simple RAG (Retrieval-Augmented Generation) application for asking questions about PDF documents.
Upload your PDF 📄, let the application process it ⚙️, and ask questions about its content 💬.
- 📤 Upload one or multiple PDF documents
- ⚙️ Asynchronous document processing
- 📑 PDF text extraction and chunking
- 🧠 Generate embeddings from document chunks
- 🔍 Semantic search using vector database
- 💬 Ask questions about uploaded documents
- 📡 Real-time document processing progress
- 📖 Built-in PDF viewer
- 📍 Navigate directly to specific PDF pages
- 🗃️ Store document embeddings and metadata in Qdrant
📄 PDF
│
▼
📤 Upload Document
│
▼
⚡ FastAPI
│
▼
📋 Redis Queue
│
▼
⚙️ Background Worker
│
┌───────┴────────┐
▼ ▼
📑 Extract Text 📡 Progress
│ │
▼ ▼
✂️ Chunking WebSocket
│ │
▼ ▼
🧠 Embedding 💻 React
│
▼
🗃️ Qdrant
│
│
💬 User Question
│
▼
🧠 Query Embedding
│
▼
🔍 Vector Search
│
▼
📚 Relevant Chunks
│
▼
🤖 LLM
│
▼
💬 AI Response
🧪 This is a learning project.
The main purpose of ask-pdf is to learn, experiment, and understand how RAG-based applications work.
This project is not intended to be production-ready.
The implementation may contain experimental approaches, simplified architecture, bugs, performance limitations, and missing production-level security or scalability considerations.