--- title: Voice-Enabled Document Chatbot using RAG emoji: 🗣️ colorFrom: blue colorTo: purple sdk: docker pinned: false license: mit --- # Voice-Enabled Document Chatbot using RAG An AI-powered web application built with Flask, LangChain, and Hugging Face that allows users to upload extensive PDF documents (100+ pages) and interact with them using both text and voice[cite: 1]. ## Features * **Retrieval-Augmented Generation (RAG):** Uses `Sentence-Transformers` (`all-MiniLM-L6-v2`) for semantic document retrieval and vector embeddings stored via Chroma[cite: 1]. * **Advanced LLM Integration:** Powered by Meta Llama (`meta-llama/Llama-3.1-8B-Instruct`) through Hugging Face Endpoints[cite: 1]. * **Voice-Enabled Interaction:** Integrated with Web Speech APIs (Speech-to-Text / Text-to-Speech) for hands-free conversational queries and audio readouts[cite: 1]. * **Responsive UI & Theming:** Features a custom Bootstrap interface with dark/light mode switching and loading animations. * **Containerized Deployment:** Packaged and deployed via Docker on Hugging Face Spaces[cite: 1]. ## Tech Stack * **Backend:** Python, Flask, Flask-CORS, LangChain, Chroma, Sentence-Transformers[cite: 1] * **AI/NLP:** Large Language Models (LLM), RAG, Hugging Face Hub, STT/TTS[cite: 1] * **Frontend:** HTML5, CSS3, JavaScript (jQuery, Bootstrap)[cite: 1] * **DevOps:** Docker, Git[cite: 1]