ChatterBox

ChatterBox is an AI-powered communication platform that allows users to chat with their friends in real-time, with AI-generated responses mimicking their friend's voice and personality. Using a short 10-second voice sample, the system clones the voice and stores embeddings for future use. Users can create personalized friend profiles, interact using voice or text, and receive responses generated by LLaMA 3.2, tailored to each friend's traits.

---

Features

Voice Cloning
Clone the voice of any person using a 10-second audio sample. Voice embeddings are saved for quicker access in future interactions.

Custom Friend Profiles
Create and manage multiple profiles with custom details:

• Name
• Personality traits
• Voice sample

Real-Time Chat
Communicate live with friends through a responsive web interface, supporting both voice and text input.

AI Responses in Friend’s Voice
Receive AI-generated replies in the cloned voice of your selected friend, enhancing the realism of conversations.

Text and Voice Messaging
Users can send messages via text or voice. The AI responds in the same mode using speech generation models.

Personality-Driven Replies
AI tailors replies to match the personality traits defined for each friend.

LLaMA 3.2 Integration
Utilizes Meta’s LLaMA 3.2 language model to generate natural, context-aware replies.

MongoDB Integration
Stores:

• User profiles
• Message history
• Audio embeddings

Efficient Voice Cloning Storage
Voice cloning is a one-time process per friend. The voice data is stored and reused to improve performance.

Advanced Voice Generation
Incorporates SpeechBrain and SpeechT5 models to produce expressive and accurate voice responses.

---

Technologies Used

• Flask (for UI and server)
• Python
• HTML, CSS, JavaScript
• MongoDB (for storing user data and voice embeddings)
• LLaMA 3.2 (AI text generation)
• SpeechBrain & SpeechT5 (voice generation)

---

Getting Started

1. Clone the Repository

git clone https://github.com/your-username/chatterbox.git cd chatterbox

2. Install dependencies

3. Download Llama 3.2 model from ollama and run locally

4. Run it using python app.py

License

• This project is licensed under the MIT License.