NovahSpeaks

Visit NovahSpeaks for a live demo.

NovahSpeaks is an AI-powered chat application designed to provide a supportive and engaging conversation experience. The application features a modern, user-friendly interface with both text and voice interaction capabilities.

📁 Project Structure

NovahSpeaks/
├── agents/           # Agent implementations
├── data/            # Core memories and profiles
├── engine/          # Core engine components
├── utils/           # Utility functions
├── main.py          # Application entry point
├── llm_config.py    # LLM configuration
└── README.md        # This file

🌟 Features

Core Features

• Interactive Chat Interface: Clean, modern UI with real-time message updates
• Voice Input: Record and transcribe your messages using the microphone
• Text-to-Speech: AI responses are converted to natural-sounding speech
• User Profiles: Personalized experience with name and emoji selection
• Memory System: Remembers past conversations and important details
• Emotional Intelligence: Analyzes and responds to emotional context
• Vocabulary Building: Tracks and learns from user's language patterns
• Personalized Communication: Tailored responses based on individual's core memories and experiences
• Context-Aware: Maintains conversation context and emotional awareness
• Simplified Language: Uses appropriate vocabulary and clear explanations
• Emotional Support: Maintains a consistent, supportive tone throughout interactions
• Memory Integration: Incorporates relevant past experiences into responses

Voice Features

• Speech-to-Text: Convert your voice to text with a simple click
• Recording Overlay: Visual feedback during voice recording with animated waves
• Stop Recording Button: Easy-to-use stop button in the recording overlay
• Audio Playback: Clear audio playback of AI responses

UI/UX Improvements

• Modern Design: Clean, intuitive interface with smooth animations
• Responsive Layout: Works seamlessly on both desktop and mobile devices
• Visual Feedback: Animated recording waves and status indicators
• Easy Navigation: Quick access to home page via the NovahSpeaks title
• Accessible Input: Large, clear text input with voice recording option
• Dull Text Box: Reduced eye strain with optimized input field colors

Technical Features

• Real-time Processing: Immediate response to user input
• Error Handling: Robust error management and recovery
• Profile Management: Automatic saving of user preferences and progress
• Memory Management: Efficient storage and retrieval of conversation history
• Emotional Analysis: Advanced sentiment and emotion detection
• Context Awareness: Maintains conversation context for relevant responses

🏗️ Architecture

The system is built on a multi-agent architecture that works together to provide personalized responses:

Core Components

1. QueryParserAgent

   • Analyzes user questions
   • Extracts intent, emotions, and key topics
   • Identifies the communication context

2. MemoryRetrievalAgent

   • Manages core memories database
   • Uses vector similarity search for relevant memories
   • Integrates with FAISS/Pinecone for efficient retrieval

3. ContextFilterAgent

   • Filters memories based on:
     • Emotional alignment
     • Vocabulary fit
     • Recency and relevance
   • Ensures appropriate cognitive load

4. ResponsePlannerAgent

   • Crafts structured prompts
   • Integrates selected memories with current context
   • Maintains consistent communication style

5. LLMResponderAgent

   • Generates responses using LLM
   • Maintains tone and style consistency
   • Ensures appropriate length and complexity

6. SanitizerAgent

   • Validates response safety
   • Ensures vocabulary compliance
   • Maintains consistent tone

🚀 Getting Started

Prerequisites

• Python 3.8+
• Groq API key
• Required Python packages (see requirements.txt)

Installation

1. Clone the repository:

git clone https://github.com/asaripa3/Novah.git
cd NovahSpeaks

2. Install dependencies:

pip install -r requirements.txt

3. Set up environment variables:

cp .env.example .env
# Add your GROQ_API_KEY to .env

4. Prepare your data:

• Add core memories to data/core_memories.jsonl
• Configure user profile in data/yahya_profile.jsonl

Usage

1. Initial Setup:

   • Enter your name
   • Choose an emoji to represent you
   • Click "Start Chatting"

2. Chatting:

   • Type your message in the text box
   • Or click the microphone icon to record your voice
   • Click the stop button in the recording overlay to end recording
   • Click "Send" to send your message

3. Voice Recording:

   • Click the microphone icon to start recording
   • A recording overlay will appear with animated waves
   • Click the "Stop Recording" button to end recording
   • Your transcribed message will appear in the text box

4. Navigation:

   • Click the "NovahSpeaks" title to return to the home page
   • Use the back button to return to previous pages