📄 PDF Question Answering Chatbot (RAG App with Telegram Bot + n8n Orchestration)

An intelligent Retrieval-Augmented Generation (RAG) application that allows users to upload PDFs, extract content, and ask contextual questions — powered by LangChain, Cohere, HuggingFace embeddings, and FAISS.
In addition to a Streamlit web app, it integrates a Telegram bot automated via n8n for seamless real-time conversational document Q&A across platforms.

---

🚀 Features

• 📤 Upload PDFs directly via Streamlit or Telegram bot
• 🧠 Chunk and embed text using HuggingFace Sentence Transformers
• 🪄 Semantic search powered by FAISS vector store
• 🤖 Answer generation using Cohere’s command-r-plus LLM through LangChain’s RAG pipeline
• 💬 Real-time streaming responses in chat-style interface
• ⚙️ n8n Orchestration:
  • Automates Telegram message reception, PDF upload handling, and API request flow
  • Cleans responses with AI Transform nodes to fix Telegram formatting issues
• 📱 Telegram Bot Integration:
  • Users can send messages or upload PDFs directly in Telegram
  • Bot forwards input to FastAPI /chat endpoint via n8n workflow
  • Receives and sends AI-generated answers instantly back to chat
• 🌐 Deployment-ready on Streamlit Cloud or Render

---

🧩 System Architecture