This project is a voice-based chatbot interface powered by OpenAI's GPT-4, using Gradio for the UI and SpeechRecognition for converting audio input to text. Users can speak into their microphone, and the chatbot will respond intelligently using conversational memory.
🎙️ Voice input through the microphone 🤖 Conversational AI using GPT-4 via LangChain 🧠 Memory buffer to retain context between interactions 🌐 Web interface using Gradio (can be shared via a link)
Install the required packages using pip:
pip install gradio openai langchain python-dotenv SpeechRecognition
Also ensure you have ffmpeg installed on your system, which is required for audio processing by Gradio.
Create a .env file in the same directory with your OpenAI API key:
OPENAI_API_KEY=your_openai_api_key_here
Launch the chatbot by running the Jupyter notebook or executing the Python code:
jupyter notebook gpt4_voicechat.ipynb
The interface will open in your browser. Speak into your microphone, and the chatbot will respond.
Gradio captures audio input. SpeechRecognition converts audio to text. LangChain feeds the recognized text to GPT-4 with memory support. The chatbot responds, and the response is displayed in the interface.
📌 Notes
If the chatbot cannot understand the audio, it will return a polite error message. The conversation memory allows GPT-4 to understand context across turns.