Voice-Powered AI Chatbot

Video Demo:

Description:

This project is a Voice-Powered AI Chatbot that allows users to interact with an AI assistant through voice input. The chatbot records the user's speech, transcribes it into text, sends it to a large language model (LLM) via the Groq API, and then speaks out the response using Text-to-Speech (TTS).

---

Features

• Real-time Audio Recording: Captures user's voice input using PyAudio (or an alternative like SoundDevice if necessary).
• Speech-to-Text (ASR): Uses a pre-trained Wav2Vec2 model from Hugging Face to transcribe the recorded speech into text.
• AI Response Generation: Sends the transcribed text to Groq API (LLM) for generating a meaningful response.
• Text-to-Speech (TTS): Converts the AI-generated response into speech using pyttsx3.
• Interruptible Response: Users can stop the chatbot from speaking by pressing Enter.
• Multi-threading: Ensures smooth execution of tasks like recording, processing, and speech output.

---

Project Structure

├── project/
│   ├── main.py                 # Main script for running the chatbot
│   ├── requirements.txt        # List of required Python dependencies
│   ├── .env                    # Stores API keys and environment variables
│   ├── output.wav              # Recorded audio file
│   ├── transcription.txt       # Transcribed text file
│   ├── README.md               # Project documentation

main.py

• Handles audio recording using either PyAudio or SoundDevice.
• Processes recorded audio and saves it in WAV format.
• Uses a speech-to-text model to transcribe audio into text.
• Interacts with the Groq API to generate AI-based responses.
• Uses TTS to read the response aloud to the user.
• Allows interruption via a keyboard event to stop speech playback.

requirements.txt

Contains all dependencies required for the project, such as:

pyaudio
sounddevice
scipy
wave
keyboard
pyttsx3
dotenv
transformers
groq

.env

Stores API keys and other environment variables.

GROQ_API_KEY=your_api_key_here

output.wav

The recorded audio file that is transcribed into text.

transcription.txt

Stores the transcribed text from the recorded audio.

---

Installation and Setup

1. Clone the Repository

git clone https://github.com/your-repo/voice-chatbot.git
cd voice-chatbot

2. Install Dependencies

pip install -r requirements.txt

3. Set Up Environment Variables

Create a .env file in the project directory and add your Groq API Key:

GROQ_API_KEY=your_api_key_here

4. Run the Chatbot

python main.py

---

Challenges and Design Decisions

• PyAudio vs SoundDevice: Initially, PyAudio had installation issues in CS50.dev, so SoundDevice was used as an alternative. However, PyAudio provided better recording stability.
• Interruptible Speech: A separate thread was created to allow users to stop the chatbot’s speech using Enter.
• Groq API Integration: The chatbot requests a longer response from the AI model for more detailed answers.
• Transcription Accuracy: Wav2Vec2 was chosen for its high accuracy in recognizing speech.

---

Future Improvements

• Improve UI/UX with a web-based interface using Flask or React.
• Add support for multiple languages in speech-to-text and text-to-speech.
• Use Whisper AI for better transcription accuracy.
• Optimize real-time response processing.

---

Conclusion

This project demonstrates how voice interaction can enhance AI-powered chatbots. By combining Speech Recognition, LLM-based text generation, and Text-to-Speech, the chatbot provides a seamless and intuitive user experience.