Urdu Voice PDF Chatbot

This project is an Urdu Voice Chatbot that allows users to upload a PDF (which can be in English) and chat with the PDF in Urdu via voice. The chatbot processes text by creating embeddings with Hugging Face Embeddings for understanding the PDF content Google Gemini is used as LLM, and converts speech to text via Streamlit's mic recorder, and provides responses in Urdu through Google Text-to-Speech (gTTS).

Features

• PDF Upload: Users can upload a PDF document.
• Urdu Voice Interaction: Users can ask questions or make inquiries about the PDF content in Urdu, and receive answers in Urdu voice.
• LangChain Integration: The project uses LangChain to manage the interactions with the language model and query the PDF content effectively.
• Hugging Face Embeddings: Embeddings are generated using Hugging Face models to provide deep semantic understanding of the uploaded PDF.
• Speech to Text: User input in Urdu is converted to text using the Streamlit mic recorder for processing.
• Text to Speech: The bot responds with text that is converted to Urdu voice using gTTS (Google Text-to-Speech).
• Google Gemini Integration: The LLM (Gemini) processes the PDF content and provides meaningful responses based on the queries.

Tech Stack

• Streamlit: For building the UI and handling voice input through the mic recorder.
• Google Gemini API: The language model used for generating coherent relevant accurate responses.
• gTTS (Google Text-to-Speech): For converting chatbot responses to Urdu voice output.
• LangChain: For efficient interaction with the language model and managing PDF queries.
• Hugging Face Embeddings: For generating embeddings of the PDF content to improve the chatbot's understanding.
• PyPDFLoader: For extracting text from the uploaded PDF files.

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/urdu-voice-pdf-chatbot.git

2. Navigate to the project directory:

   cd urdu-voice-pdf-chatbot

3. Install the required dependencies:

   pip install -r requirements.txt

4. Add your Google Gemini API key in the .env file:

   GEMINI_API_KEY=your-api-key-here

Usage

1. Run the application:

   streamlit run app.py

2. Upload a PDF document (English).

3. Use the microphone to ask questions in Urdu.

4. The chatbot will respond in Urdu voice, explaining the content or answering questions based on the PDF.

Author: Arifa Batool

Live Application: https://urdu-voice-pdf-chatbot.streamlit.app/