the-voice-chatbot

the-voice-chatbot is a Streamlit-based web application that allows users to chat with a virtual voice teacher. Powered by LangChain and OpenAI, this enables users to talk to a virtual voice teacher that has been trained with voice-specific resources.

Features

• Natural Language Queries: Ask questions about voice and vocal training.
• Chat History: Maintains a conversation history for context-aware responses.
• Source Documents: Displays the sections of the PDF used to generate answers.
• OpenAI Integration: Uses OpenAI's GPT models for generating responses.
• Streamlit UI: Simple and intuitive user interface.

Demo

How It Works

1. PDF Processing: The app uses PyPDFLoader to extract text from the uploaded PDF.
2. Text Splitting: The text is split into smaller chunks using RecursiveCharacterTextSplitter.
3. Vector Database: A vector database is created using DocArrayInMemorySearch and OpenAI embeddings.
4. Chat Interface: Users can ask questions, and the app retrieves relevant information from the vector database using LangChain's ConversationalRetrievalChain.
5. Response Generation: OpenAI's GPT model generates responses based on the retrieved information.

Installation

Follow these steps to set up and run the project locally.

Prerequisites

• Python 3.8 or higher
• An OpenAI API key (get it from OpenAI)

Steps

Go to https://the-voice-chatbot.streamlit.app/