Merge branch 'main' of https://github.com/GamerBoss101/HooHacks-12

.gitignore

@@ -134,3 +134,4 @@ dist
 .yarn/build-state.yml
 .yarn/install-state.gz
 .pnp.*
+Backend/test.py

Backend/server.py

@@ -1,297 +1,699 @@
 import os
 import base64
 import json
-import asyncio
+import time
+import math
+import gc
+import logging
+import numpy as np
 import torch
 import torchaudio
-import numpy as np
 from io import BytesIO
 from typing import List, Dict, Any, Optional
-from fastapi import FastAPI, WebSocket, WebSocketDisconnect
+from flask import Flask, request, send_from_directory, Response
-from fastapi.middleware.cors import CORSMiddleware
+from flask_cors import CORS
-from pydantic import BaseModel
+from flask_socketio import SocketIO, emit, disconnect
 from generator import load_csm_1b, Segment
-import uvicorn
-import time
 from collections import deque
+from threading import Lock
 
-# Select device
+# Configure logging
-if torch.cuda.is_available():
+logging.basicConfig(
-    device = "cuda"
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
+)
+logger = logging.getLogger("sesame-server")
+
+# Determine best compute device
+if torch.backends.mps.is_available():
+    device = "mps"
+elif torch.cuda.is_available():
+    try:
+        # Test CUDA functionality
+        torch.rand(10, device="cuda")
+        torch.backends.cuda.matmul.allow_tf32 = True
+        torch.backends.cudnn.allow_tf32 = True
+        torch.backends.cudnn.benchmark = True
+        device = "cuda"
+        logger.info("CUDA is fully functional")
+    except Exception as e:
+        logger.warning(f"CUDA available but not working correctly: {e}")
+        device = "cpu"
 else:
     device = "cpu"
-print(f"Using device: {device}")
+    logger.info("Using CPU")
 
-# Initialize the model
+# Constants and Configuration
-generator = load_csm_1b(device=device)
+SILENCE_THRESHOLD = 0.01
+SILENCE_DURATION_SEC = 0.75
+MAX_BUFFER_SIZE = 30  # Maximum chunks to buffer before processing
+CHUNK_SIZE_MS = 500  # Size of audio chunks when streaming responses
 
-app = FastAPI()
+# Define the base directory and static files directory
+base_dir = os.path.dirname(os.path.abspath(__file__))
+static_dir = os.path.join(base_dir, "static")
+os.makedirs(static_dir, exist_ok=True)
 
-# Add CORS middleware to allow cross-origin requests
+# Define a simple energy-based speech detector
-app.add_middleware(
+class SpeechDetector:
-    CORSMiddleware,
-    allow_origins=["*"],  # Allow all origins in development
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-# Connection manager to handle multiple clients
-class ConnectionManager:
     def __init__(self):
-        self.active_connections: List[WebSocket] = []
+        self.min_speech_energy = 0.01
+        self.speech_window = 0.2  # seconds
+    
+    def detect_speech(self, audio_tensor, sample_rate):
+        # Calculate frame size based on window size
+        frame_size = int(sample_rate * self.speech_window)
+        
+        # If audio is shorter than frame size, use the entire audio
+        if audio_tensor.shape[0] < frame_size:
+            frames = [audio_tensor]
+        else:
+            # Split audio into frames
+            frames = [audio_tensor[i:i+frame_size] for i in range(0, len(audio_tensor), frame_size)]
+        
+        # Calculate energy per frame
+        energies = [torch.mean(frame**2).item() for frame in frames]
+        
+        # Determine if there's speech based on energy threshold
+        has_speech = any(e > self.min_speech_energy for e in energies)
+        
+        return has_speech
 
-    async def connect(self, websocket: WebSocket):
+speech_detector = SpeechDetector()
-        await websocket.accept()
+logger.info("Initialized simple speech detector")
-        self.active_connections.append(websocket)
 
-    def disconnect(self, websocket: WebSocket):
+# Model Loading Functions
-        self.active_connections.remove(websocket)
+def load_speech_models():
-
+    """Load speech generation model"""
-manager = ConnectionManager()
+    # Load speech generation model (Sesame CSM)
-
-# Silence detection parameters
-SILENCE_THRESHOLD = 0.01  # Adjust based on your audio normalization
-SILENCE_DURATION_SEC = 1.0  # How long silence must persist to be considered "stopped talking"
-
-# Helper function to convert audio data
-async def decode_audio_data(audio_data: str) -> torch.Tensor:
-    """Decode base64 audio data to a torch tensor"""
     try:
+        logger.info(f"Loading Sesame CSM model on {device}...")
+        generator = load_csm_1b(device=device)
+        logger.info("Sesame CSM model loaded successfully")
+    except Exception as e:
+        logger.error(f"Error loading Sesame CSM on {device}: {e}")
+        if device == "cuda":
+            try:
+                logger.info("Trying to load Sesame CSM on CPU instead...")
+                generator = load_csm_1b(device="cpu")
+                logger.info("Sesame CSM model loaded on CPU successfully")
+            except Exception as cpu_error:
+                logger.critical(f"Failed to load speech synthesis model: {cpu_error}")
+                raise RuntimeError("Failed to load speech synthesis model")
+        else:
+            raise RuntimeError("Failed to load speech synthesis model on any device")
+    
+    return generator
+
+# Load speech model
+generator = load_speech_models()
+
+# Set up Flask and Socket.IO
+app = Flask(__name__)
+CORS(app)
+socketio = SocketIO(app, cors_allowed_origins="*", async_mode='eventlet')
+
+# Socket connection management
+thread_lock = Lock()
+active_clients = {}  # Map client_id to client context
+
+# Audio Utility Functions
+def decode_audio_data(audio_data: str) -> torch.Tensor:
+    """Decode base64 audio data to a torch tensor with improved error handling"""
+    try:
+        # Skip empty audio data
+        if not audio_data or len(audio_data) < 100:
+            logger.warning("Empty or too short audio data received")
+            return torch.zeros(generator.sample_rate // 2)  # 0.5 seconds of silence
+            
+        # Extract the actual base64 content
+        if ',' in audio_data:
+            audio_data = audio_data.split(',')[1]
+            
         # Decode base64 audio data
-        binary_data = base64.b64decode(audio_data.split(',')[1] if ',' in audio_data else audio_data)
+        try:
+            binary_data = base64.b64decode(audio_data)
+            logger.debug(f"Decoded base64 data: {len(binary_data)} bytes")
+            
+            # Check if we have enough data for a valid WAV
+            if len(binary_data) < 44:  # WAV header is 44 bytes
+                logger.warning("Data too small to be a valid WAV file")
+                return torch.zeros(generator.sample_rate // 2)
+        except Exception as e:
+            logger.error(f"Base64 decoding error: {e}")
+            return torch.zeros(generator.sample_rate // 2)
         
-        # Save to a temporary WAV file first
+        # Multiple approaches to handle audio data
-        temp_file = BytesIO(binary_data)
+        audio_tensor = None
+        sample_rate = None
         
-        # Load audio from binary data, explicitly specifying the format
+        # Approach 1: Direct loading with torchaudio
-        audio_tensor, sample_rate = torchaudio.load(temp_file, format="wav")
+        try:
+            with BytesIO(binary_data) as temp_file:
+                temp_file.seek(0)
+                audio_tensor, sample_rate = torchaudio.load(temp_file, format="wav")
+                logger.debug(f"Loaded audio: shape={audio_tensor.shape}, rate={sample_rate}Hz")
+                
+                # Validate tensor
+                if audio_tensor.numel() == 0 or torch.isnan(audio_tensor).any():
+                    raise ValueError("Invalid audio tensor")
+        except Exception as e:
+            logger.warning(f"Direct loading failed: {e}")
+            
+            # Approach 2: Using wave module and numpy
+            try:
+                temp_path = os.path.join(base_dir, f"temp_{time.time()}.wav")
+                with open(temp_path, 'wb') as f:
+                    f.write(binary_data)
+                
+                import wave
+                with wave.open(temp_path, 'rb') as wf:
+                    n_channels = wf.getnchannels()
+                    sample_width = wf.getsampwidth()
+                    sample_rate = wf.getframerate()
+                    n_frames = wf.getnframes()
+                    frames = wf.readframes(n_frames)
+                    
+                    # Convert to numpy array
+                    if sample_width == 2:  # 16-bit audio
+                        data = np.frombuffer(frames, dtype=np.int16).astype(np.float32) / 32768.0
+                    elif sample_width == 1:  # 8-bit audio
+                        data = np.frombuffer(frames, dtype=np.uint8).astype(np.float32) / 128.0 - 1.0
+                    else:
+                        raise ValueError(f"Unsupported sample width: {sample_width}")
+                    
+                    # Convert to mono if needed
+                    if n_channels > 1:
+                        data = data.reshape(-1, n_channels)
+                        data = data.mean(axis=1)
+                    
+                    # Convert to torch tensor
+                    audio_tensor = torch.from_numpy(data)
+                    logger.info(f"Loaded audio using wave: shape={audio_tensor.shape}")
+                
+                # Clean up temp file
+                if os.path.exists(temp_path):
+                    os.remove(temp_path)
+                    
+            except Exception as e2:
+                logger.error(f"All audio loading methods failed: {e2}")
+                return torch.zeros(generator.sample_rate // 2)
         
+        # Format corrections
+        if audio_tensor is None:
+            return torch.zeros(generator.sample_rate // 2)
+            
+        # Ensure audio is mono
+        if len(audio_tensor.shape) > 1 and audio_tensor.shape[0] > 1:
+            audio_tensor = torch.mean(audio_tensor, dim=0)
+        
+        # Ensure 1D tensor
+        audio_tensor = audio_tensor.squeeze()
+            
         # Resample if needed
         if sample_rate != generator.sample_rate:
-            audio_tensor = torchaudio.functional.resample(
+            try:
-                audio_tensor.squeeze(0), 
+                logger.debug(f"Resampling from {sample_rate}Hz to {generator.sample_rate}Hz")
-                orig_freq=sample_rate, 
+                resampler = torchaudio.transforms.Resample(
-                new_freq=generator.sample_rate
+                    orig_freq=sample_rate, 
-            )
+                    new_freq=generator.sample_rate
-        else:
+                )
-            audio_tensor = audio_tensor.squeeze(0)
+                audio_tensor = resampler(audio_tensor)
-            
+            except Exception as e:
+                logger.warning(f"Resampling error: {e}")
+        
+        # Normalize audio to avoid issues
+        if torch.abs(audio_tensor).max() > 0:
+            audio_tensor = audio_tensor / torch.abs(audio_tensor).max()
+        
         return audio_tensor
     except Exception as e:
-        print(f"Error decoding audio: {str(e)}")
+        logger.error(f"Unhandled error in decode_audio_data: {e}")
-        # Return a small silent audio segment as fallback
+        return torch.zeros(generator.sample_rate // 2)
-        return torch.zeros(generator.sample_rate // 2)  # 0.5 seconds of silence
 
-
+def encode_audio_data(audio_tensor: torch.Tensor) -> str:
-async def encode_audio_data(audio_tensor: torch.Tensor) -> str:
     """Encode torch tensor audio to base64 string"""
-    buf = BytesIO()
+    try:
-    torchaudio.save(buf, audio_tensor.unsqueeze(0).cpu(), generator.sample_rate, format="wav")
+        buf = BytesIO()
-    buf.seek(0)
+        torchaudio.save(buf, audio_tensor.unsqueeze(0).cpu(), generator.sample_rate, format="wav")
-    audio_base64 = base64.b64encode(buf.read()).decode('utf-8')
+        buf.seek(0)
-    return f"data:audio/wav;base64,{audio_base64}"
+        audio_base64 = base64.b64encode(buf.read()).decode('utf-8')
+        return f"data:audio/wav;base64,{audio_base64}"
+    except Exception as e:
+        logger.error(f"Error encoding audio: {e}")
+        # Return a minimal silent audio file
+        silence = torch.zeros(generator.sample_rate // 2).unsqueeze(0)
+        buf = BytesIO()
+        torchaudio.save(buf, silence, generator.sample_rate, format="wav")
+        buf.seek(0)
+        return f"data:audio/wav;base64,{base64.b64encode(buf.read()).decode('utf-8')}"
 
-
+def process_speech(audio_tensor: torch.Tensor, client_id: str) -> str:
-@app.websocket("/ws")
+    """Process speech and return a simple response"""
-async def websocket_endpoint(websocket: WebSocket):
+    # In this simplified version, we'll just check if there's sound
-    await manager.connect(websocket)
+    # and provide basic responses instead of doing actual speech recognition
-    context_segments = []  # Store conversation context
-    streaming_buffer = []  # Buffer for streaming audio chunks
-    is_streaming = False
     
-    # Variables for silence detection
+    if speech_detector and speech_detector.detect_speech(audio_tensor, generator.sample_rate):
-    last_active_time = time.time()
+        # Generate a response based on audio energy
-    is_silence = False
+        energy = torch.mean(torch.abs(audio_tensor)).item()
-    energy_window = deque(maxlen=10)  # For tracking recent audio energy
+        
+        if energy > 0.1:  # Louder speech
+            return "I heard you speaking clearly. How can I help you today?"
+        elif energy > 0.05:  # Moderate speech
+            return "I heard you say something. Could you please repeat that?"
+        else:  # Soft speech
+            return "I detected some speech, but it was quite soft. Could you speak up a bit?"
+    else:
+        return "I didn't detect any speech. Could you please try again?"
+
+def generate_response(text: str, conversation_history: List[Segment]) -> str:
+    """Generate a contextual response based on the transcribed text"""
+    # Simple response logic - can be replaced with a more sophisticated LLM
+    responses = {
+        "hello": "Hello there! How can I help you today?",
+        "hi": "Hi there! What can I do for you?",
+        "how are you": "I'm doing well, thanks for asking! How about you?",
+        "what is your name": "I'm Sesame, your voice assistant. How can I help you?",
+        "who are you": "I'm Sesame, an AI voice assistant. I'm here to chat with you!",
+        "bye": "Goodbye! It was nice chatting with you.",
+        "thank you": "You're welcome! Is there anything else I can help with?",
+        "weather": "I don't have real-time weather data, but I hope it's nice where you are!",
+        "help": "I can chat with you using natural voice. Just speak normally and I'll respond.",
+        "what can you do": "I can have a conversation with you, answer questions, and provide assistance with various topics.",
+    }
+    
+    text_lower = text.lower()
+    
+    # Check for matching keywords
+    for key, response in responses.items():
+        if key in text_lower:
+            return response
+    
+    # Default responses based on text length
+    if not text:
+        return "I didn't catch that. Could you please repeat?"
+    elif len(text) < 10:
+        return "Thanks for your message. Could you elaborate a bit more?"
+    else:
+        return f"I heard you speaking. That's interesting! Can you tell me more about that?"
+
+# Flask Routes
+@app.route('/')
+def index():
+    return send_from_directory(base_dir, 'index.html')
+
+@app.route('/favicon.ico')
+def favicon():
+    if os.path.exists(os.path.join(static_dir, 'favicon.ico')):
+        return send_from_directory(static_dir, 'favicon.ico')
+    return Response(status=204)
+
+@app.route('/voice-chat.js')
+def voice_chat_js():
+    return send_from_directory(base_dir, 'voice-chat.js')
+
+@app.route('/static/<path:path>')
+def serve_static(path):
+    return send_from_directory(static_dir, path)
+
+# Socket.IO Event Handlers
+@socketio.on('connect')
+def handle_connect():
+    client_id = request.sid
+    logger.info(f"Client connected: {client_id}")
+    
+    # Initialize client context
+    active_clients[client_id] = {
+        'context_segments': [],
+        'streaming_buffer': [],
+        'is_streaming': False,
+        'is_silence': False,
+        'last_active_time': time.time(),
+        'energy_window': deque(maxlen=10)
+    }
+    
+    emit('status', {'type': 'connected', 'message': 'Connected to server'})
+
+@socketio.on('disconnect')
+def handle_disconnect():
+    client_id = request.sid
+    if client_id in active_clients:
+        del active_clients[client_id]
+    logger.info(f"Client disconnected: {client_id}")
+
+@socketio.on('generate')
+def handle_generate(data):
+    client_id = request.sid
+    if client_id not in active_clients:
+        emit('error', {'message': 'Client not registered'})
+        return
     
     try:
-        while True:
+        text = data.get('text', '')
-            # Receive JSON data from client
+        speaker_id = data.get('speaker', 0)
-            data = await websocket.receive_text()
+        
-            request = json.loads(data)
+        logger.info(f"Generating audio for: '{text}' with speaker {speaker_id}")
+        
+        # Generate audio response
+        audio_tensor = generator.generate(
+            text=text,
+            speaker=speaker_id,
+            context=active_clients[client_id]['context_segments'],
+            max_audio_length_ms=10_000,
+        )
+        
+        # Add to conversation context
+        active_clients[client_id]['context_segments'].append(
+            Segment(text=text, speaker=speaker_id, audio=audio_tensor)
+        )
+        
+        # Convert audio to base64 and send back to client
+        audio_base64 = encode_audio_data(audio_tensor)
+        emit('audio_response', {
+            'type': 'audio_response',
+            'audio': audio_base64,
+            'text': text
+        })
+        
+    except Exception as e:
+        logger.error(f"Error generating audio: {e}")
+        emit('error', {
+            'type': 'error',
+            'message': f"Error generating audio: {str(e)}"
+        })
+
+@socketio.on('add_to_context')
+def handle_add_to_context(data):
+    client_id = request.sid
+    if client_id not in active_clients:
+        emit('error', {'message': 'Client not registered'})
+        return
+    
+    try:
+        text = data.get('text', '')
+        speaker_id = data.get('speaker', 0)
+        audio_data = data.get('audio', '')
+        
+        # Convert received audio to tensor
+        audio_tensor = decode_audio_data(audio_data)
+        
+        # Add to conversation context
+        active_clients[client_id]['context_segments'].append(
+            Segment(text=text, speaker=speaker_id, audio=audio_tensor)
+        )
+        
+        emit('context_updated', {
+            'type': 'context_updated',
+            'message': 'Audio added to context'
+        })
+        
+    except Exception as e:
+        logger.error(f"Error adding to context: {e}")
+        emit('error', {
+            'type': 'error',
+            'message': f"Error processing audio: {str(e)}"
+        })
+
+@socketio.on('clear_context')
+def handle_clear_context():
+    client_id = request.sid
+    if client_id in active_clients:
+        active_clients[client_id]['context_segments'] = []
+        
+    emit('context_updated', {
+        'type': 'context_updated',
+        'message': 'Context cleared'
+    })
+
+@socketio.on('stream_audio')
+def handle_stream_audio(data):
+    client_id = request.sid
+    if client_id not in active_clients:
+        emit('error', {'message': 'Client not registered'})
+        return
+    
+    client = active_clients[client_id]
+    
+    try:
+        speaker_id = data.get('speaker', 0)
+        audio_data = data.get('audio', '')
+        
+        # Skip if no audio data (might be just a connection test)
+        if not audio_data:
+            logger.debug("Empty audio data received, ignoring")
+            return
+        
+        # Convert received audio to tensor
+        audio_chunk = decode_audio_data(audio_data)
+        
+        # Start streaming mode if not already started
+        if not client['is_streaming']:
+            client['is_streaming'] = True
+            client['streaming_buffer'] = []
+            client['energy_window'].clear()
+            client['is_silence'] = False
+            client['last_active_time'] = time.time()
+            logger.info(f"[{client_id[:8]}] Streaming started with speaker ID: {speaker_id}")
+            emit('streaming_status', {
+                'type': 'streaming_status',
+                'status': 'started'
+            })
+        
+        # Calculate audio energy for silence detection
+        chunk_energy = torch.mean(torch.abs(audio_chunk)).item()
+        client['energy_window'].append(chunk_energy)
+        avg_energy = sum(client['energy_window']) / len(client['energy_window'])
+        
+        # Check if audio is silent
+        current_silence = avg_energy < SILENCE_THRESHOLD
+        
+        # Track silence transition
+        if not client['is_silence'] and current_silence:
+            # Transition to silence
+            client['is_silence'] = True
+            client['last_active_time'] = time.time()
+        elif client['is_silence'] and not current_silence:
+            # User started talking again
+            client['is_silence'] = False
+        
+        # Add chunk to buffer regardless of silence state
+        client['streaming_buffer'].append(audio_chunk)
             
-            action = request.get("action")
+        # Check if silence has persisted long enough to consider "stopped talking"
+        silence_elapsed = time.time() - client['last_active_time']
+        
+        if client['is_silence'] and silence_elapsed >= SILENCE_DURATION_SEC and len(client['streaming_buffer']) > 0:
+            # User has stopped talking - process the collected audio
+            logger.info(f"[{client_id[:8]}] Processing audio after {silence_elapsed:.2f}s of silence")
+            process_complete_utterance(client_id, client, speaker_id)
+        
+        # If buffer gets too large without silence, process it anyway
+        elif len(client['streaming_buffer']) >= MAX_BUFFER_SIZE:
+            logger.info(f"[{client_id[:8]}] Processing long audio segment without silence")
+            process_complete_utterance(client_id, client, speaker_id, is_incomplete=True)
             
-            if action == "generate":
+            # Keep half of the buffer for context (sliding window approach)
+            half_point = len(client['streaming_buffer']) // 2
+            client['streaming_buffer'] = client['streaming_buffer'][half_point:]
+            
+    except Exception as e:
+        import traceback
+        traceback.print_exc()
+        logger.error(f"Error processing streaming audio: {e}")
+        emit('error', {
+            'type': 'error',
+            'message': f"Error processing streaming audio: {str(e)}"
+        })
+
+def process_complete_utterance(client_id, client, speaker_id, is_incomplete=False):
+    """Process a complete utterance (after silence or buffer limit)"""
+    try:
+        # Combine audio chunks
+        full_audio = torch.cat(client['streaming_buffer'], dim=0)
+        
+        # Process audio to generate a response (no speech recognition)
+        generated_text = process_speech(full_audio, client_id)
+        
+        # Add suffix for incomplete utterances
+        if is_incomplete:
+            generated_text += " (processing continued speech...)"
+        
+        # Log the generated text
+        logger.info(f"[{client_id[:8]}] Generated text: '{generated_text}'")
+        
+        # Handle the result
+        if generated_text:
+            # Add user message to context
+            user_segment = Segment(text=generated_text, speaker=speaker_id, audio=full_audio)
+            client['context_segments'].append(user_segment)
+            
+            # Send the text to client
+            emit('transcription', {
+                'type': 'transcription',
+                'text': generated_text
+            }, room=client_id)
+            
+            # Only generate a response if this is a complete utterance
+            if not is_incomplete:
+                # Generate a contextual response
+                response_text = generate_response(generated_text, client['context_segments'])
+                logger.info(f"[{client_id[:8]}] Generating response: '{response_text}'")
+                
+                # Let the client know we're processing
+                emit('processing_status', {
+                    'type': 'processing_status',
+                    'status': 'generating_audio',
+                    'message': 'Generating audio response...'
+                }, room=client_id)
+                
+                # Generate audio for the response
                 try:
-                    text = request.get("text", "")
+                    # Use a different speaker than the user
-                    speaker_id = request.get("speaker", 0)
+                    ai_speaker_id = 1 if speaker_id == 0 else 0
                     
-                    # Generate audio response
+                    # Generate the full response
-                    print(f"Generating audio for: '{text}' with speaker {speaker_id}")
                     audio_tensor = generator.generate(
-                        text=text,
+                        text=response_text,
-                        speaker=speaker_id,
+                        speaker=ai_speaker_id,
-                        context=context_segments,
+                        context=client['context_segments'],
                         max_audio_length_ms=10_000,
                     )
                     
-                    # Add to conversation context
+                    # Add response to context
-                    context_segments.append(Segment(text=text, speaker=speaker_id, audio=audio_tensor))
+                    ai_segment = Segment(
+                        text=response_text, 
+                        speaker=ai_speaker_id, 
+                        audio=audio_tensor
+                    )
+                    client['context_segments'].append(ai_segment)
                     
                     # Convert audio to base64 and send back to client
-                    audio_base64 = await encode_audio_data(audio_tensor)
+                    audio_base64 = encode_audio_data(audio_tensor)
-                    await websocket.send_json({
+                    emit('audio_response', {
-                        "type": "audio_response",
+                        'type': 'audio_response',
-                        "audio": audio_base64
+                        'text': response_text,
-                    })
+                        'audio': audio_base64
+                    }, room=client_id)
+                    
+                    logger.info(f"[{client_id[:8]}] Audio response sent")
+                    
                 except Exception as e:
-                    print(f"Error generating audio: {str(e)}")
+                    logger.error(f"Error generating audio response: {e}")
-                    await websocket.send_json({
+                    emit('error', {
-                        "type": "error",
+                        'type': 'error',
-                        "message": f"Error generating audio: {str(e)}"
+                        'message': "Sorry, there was an error generating the audio response."
-                    })
+                    }, room=client_id)
-                
+        else:
-            elif action == "add_to_context":
+            # If processing failed, send a notification
-                try:
+            emit('error', {
-                    text = request.get("text", "")
+                'type': 'error',
-                    speaker_id = request.get("speaker", 0)
+                'message': "Sorry, I couldn't understand what you said. Could you try again?"
-                    audio_data = request.get("audio", "")
+            }, room=client_id)
-                    
+        
-                    # Convert received audio to tensor
+        # Only clear buffer for complete utterances
-                    audio_tensor = await decode_audio_data(audio_data)
+        if not is_incomplete:
-                    
+            # Reset state
-                    # Add to conversation context
+            client['streaming_buffer'] = []
-                    context_segments.append(Segment(text=text, speaker=speaker_id, audio=audio_tensor))
+            client['energy_window'].clear()
-                    
+            client['is_silence'] = False
-                    await websocket.send_json({
+            client['last_active_time'] = time.time()
-                        "type": "context_updated",
-                        "message": "Audio added to context"
-                    })
-                except Exception as e:
-                    print(f"Error adding to context: {str(e)}")
-                    await websocket.send_json({
-                        "type": "error",
-                        "message": f"Error processing audio: {str(e)}"
-                    })
-                
-            elif action == "clear_context":
-                context_segments = []
-                await websocket.send_json({
-                    "type": "context_updated",
-                    "message": "Context cleared"
-                })
             
-            elif action == "stream_audio":
-                try:
-                    speaker_id = request.get("speaker", 0)
-                    audio_data = request.get("audio", "")
-                    
-                    # Convert received audio to tensor
-                    audio_chunk = await decode_audio_data(audio_data)
-                    
-                    # Start streaming mode if not already started
-                    if not is_streaming:
-                        is_streaming = True
-                        streaming_buffer = []
-                        energy_window.clear()
-                        is_silence = False
-                        last_active_time = time.time()
-                        await websocket.send_json({
-                            "type": "streaming_status",
-                            "status": "started"
-                        })
-                    
-                    # Calculate audio energy for silence detection
-                    chunk_energy = torch.mean(torch.abs(audio_chunk)).item()
-                    energy_window.append(chunk_energy)
-                    avg_energy = sum(energy_window) / len(energy_window)
-                    
-                    # Check if audio is silent
-                    current_silence = avg_energy < SILENCE_THRESHOLD
-                    
-                    # Track silence transition
-                    if not is_silence and current_silence:
-                        # Transition to silence
-                        is_silence = True
-                        last_active_time = time.time()
-                    elif is_silence and not current_silence:
-                        # User started talking again
-                        is_silence = False
-                    
-                    # Add chunk to buffer regardless of silence state
-                    streaming_buffer.append(audio_chunk)
-                    
-                    # Check if silence has persisted long enough to consider "stopped talking"
-                    silence_elapsed = time.time() - last_active_time
-                    
-                    if is_silence and silence_elapsed >= SILENCE_DURATION_SEC and len(streaming_buffer) > 0:
-                        # User has stopped talking - process the collected audio
-                        full_audio = torch.cat(streaming_buffer, dim=0)
-                        
-                        # Process with speech-to-text (you would need to implement this)
-                        # For now, just use a placeholder text
-                        text = f"User audio from speaker {speaker_id}"
-                        
-                        print(f"Detected end of speech, processing {len(streaming_buffer)} chunks")
-                        
-                        # Add to conversation context
-                        context_segments.append(Segment(text=text, speaker=speaker_id, audio=full_audio))
-                        
-                        # Generate response
-                        response_text = "This is a response to what you just said"
-                        audio_tensor = generator.generate(
-                            text=response_text,
-                            speaker=1 if speaker_id == 0 else 0,  # Use opposite speaker
-                            context=context_segments,
-                            max_audio_length_ms=10_000,
-                        )
-                        
-                        # Convert audio to base64 and send back to client
-                        audio_base64 = await encode_audio_data(audio_tensor)
-                        await websocket.send_json({
-                            "type": "audio_response",
-                            "audio": audio_base64
-                        })
-                        
-                        # Clear buffer and reset silence detection
-                        streaming_buffer = []
-                        energy_window.clear()
-                        is_silence = False
-                        last_active_time = time.time()
-                    
-                    # If buffer gets too large without silence, process it anyway
-                    # This prevents memory issues with very long streams
-                    elif len(streaming_buffer) >= 30:  # ~6 seconds of audio at 5 chunks/sec
-                        print("Buffer limit reached, processing audio")
-                        full_audio = torch.cat(streaming_buffer, dim=0)
-                        text = f"Continued speech from speaker {speaker_id}"
-                        context_segments.append(Segment(text=text, speaker=speaker_id, audio=full_audio))
-                        streaming_buffer = []
-                        
-                except Exception as e:
-                    print(f"Error processing streaming audio: {str(e)}")
-                    await websocket.send_json({
-                        "type": "error",
-                        "message": f"Error processing streaming audio: {str(e)}"
-                    })
-            
-            elif action == "stop_streaming":
-                is_streaming = False
-                if streaming_buffer:
-                    # Process any remaining audio in the buffer
-                    full_audio = torch.cat(streaming_buffer, dim=0)
-                    text = f"Final streaming audio from speaker {request.get('speaker', 0)}"
-                    context_segments.append(Segment(text=text, speaker=request.get("speaker", 0), audio=full_audio))
-                
-                streaming_buffer = []
-                await websocket.send_json({
-                    "type": "streaming_status",
-                    "status": "stopped"
-                })
-    
-    except WebSocketDisconnect:
-        manager.disconnect(websocket)
-        print("Client disconnected")
     except Exception as e:
-        print(f"Error: {str(e)}")
+        logger.error(f"Error processing utterance: {e}")
-        await websocket.send_json({
+        emit('error', {
-            "type": "error",
+            'type': 'error',
-            "message": str(e)
+            'message': f"Error processing audio: {str(e)}"
-        })
+        }, room=client_id)
-        manager.disconnect(websocket)
 
+@socketio.on('stop_streaming')
+def handle_stop_streaming(data):
+    client_id = request.sid
+    if client_id not in active_clients:
+        return
+    
+    client = active_clients[client_id]
+    client['is_streaming'] = False
+    
+    if client['streaming_buffer'] and len(client['streaming_buffer']) > 5:
+        # Process any remaining audio in the buffer
+        logger.info(f"[{client_id[:8]}] Processing final audio buffer on stop")
+        process_complete_utterance(client_id, client, data.get("speaker", 0))
+    
+    client['streaming_buffer'] = []
+    emit('streaming_status', {
+        'type': 'streaming_status',
+        'status': 'stopped'
+    })
 
+def stream_audio_to_client(client_id, audio_tensor, text, speaker_id, chunk_size_ms=CHUNK_SIZE_MS):
+    """Stream audio to client in chunks to simulate real-time generation"""
+    try:
+        if client_id not in active_clients:
+            logger.warning(f"Client {client_id} not found for streaming")
+            return
+            
+        # Calculate chunk size in samples
+        chunk_size = int(generator.sample_rate * chunk_size_ms / 1000)
+        total_chunks = math.ceil(audio_tensor.size(0) / chunk_size)
+        
+        logger.info(f"Streaming audio in {total_chunks} chunks of {chunk_size_ms}ms each")
+        
+        # Send initial response with text but no audio yet
+        socketio.emit('audio_response_start', {
+            'type': 'audio_response_start',
+            'text': text,
+            'total_chunks': total_chunks
+        }, room=client_id)
+        
+        # Stream each chunk
+        for i in range(total_chunks):
+            start_idx = i * chunk_size
+            end_idx = min(start_idx + chunk_size, audio_tensor.size(0))
+            
+            # Extract chunk
+            chunk = audio_tensor[start_idx:end_idx]
+            
+            # Encode chunk
+            chunk_base64 = encode_audio_data(chunk)
+            
+            # Send chunk
+            socketio.emit('audio_response_chunk', {
+                'type': 'audio_response_chunk',
+                'chunk_index': i,
+                'total_chunks': total_chunks,
+                'audio': chunk_base64,
+                'is_last': i == total_chunks - 1
+            }, room=client_id)
+            
+            # Brief pause between chunks to simulate streaming
+            time.sleep(0.1)
+            
+        # Send completion message
+        socketio.emit('audio_response_complete', {
+            'type': 'audio_response_complete',
+            'text': text
+        }, room=client_id)
+        
+        logger.info(f"Audio streaming complete: {total_chunks} chunks sent")
+        
+    except Exception as e:
+        logger.error(f"Error streaming audio to client: {e}")
+        import traceback
+        traceback.print_exc()
+
+# Main server start
 if __name__ == "__main__":
-    uvicorn.run(app, host="localhost", port=8000)
+    print(f"\n{'='*60}")
+    print(f"🔊 Sesame AI Voice Chat Server")
+    print(f"{'='*60}")
+    print(f"📡 Server Information:")
+    print(f"   - Local URL: http://localhost:5000")
+    print(f"   - Network URL: http://<your-ip-address>:5000")
+    print(f"{'='*60}")
+    print(f"🌐 Device: {device.upper()}")
+    print(f"🧠 Models: Sesame CSM (TTS only)")
+    print(f"🔧 Serving from: {os.path.join(base_dir, 'index.html')}")
+    print(f"{'='*60}")
+    print(f"Ready to receive connections! Press Ctrl+C to stop the server.\n")
+    
+    socketio.run(app, host="0.0.0.0", port=5000, debug=False)

Backend/voice-chat.js

@@ -0,0 +1,852 @@
+/**
+ * Sesame AI Voice Chat Client
+ * 
+ * A web client that connects to a Sesame AI voice chat server and enables 
+ * real-time voice conversation with an AI assistant.
+ */
+
+// Configuration constants
+const SERVER_URL = window.location.hostname === 'localhost' ? 
+    'http://localhost:5000' : window.location.origin;
+const ENERGY_WINDOW_SIZE = 15;
+const CLIENT_SILENCE_DURATION_MS = 750;
+
+// DOM elements
+const elements = {
+    conversation: null,
+    streamButton: null,
+    clearButton: null,
+    thresholdSlider: null,
+    thresholdValue: null,
+    visualizerCanvas: null,
+    visualizerLabel: null,
+    volumeLevel: null,
+    statusDot: null,
+    statusText: null,
+    speakerSelection: null,
+    autoPlayResponses: null,
+    showVisualizer: null
+};
+
+// Application state
+const state = {
+    socket: null,
+    audioContext: null,
+    analyser: null,
+    microphone: null,
+    streamProcessor: null,
+    isStreaming: false,
+    isSpeaking: false,
+    silenceThreshold: 0.01,
+    energyWindow: [],
+    silenceTimer: null,
+    volumeUpdateInterval: null,
+    visualizerAnimationFrame: null,
+    currentSpeaker: 0
+};
+
+// Visualizer variables
+let canvasContext = null;
+let visualizerBufferLength = 0;
+let visualizerDataArray = null;
+
+// Initialize the application
+function initializeApp() {
+    // Initialize the UI elements
+    initializeUIElements();
+    
+    // Initialize socket.io connection
+    setupSocketConnection();
+    
+    // Setup event listeners
+    setupEventListeners();
+    
+    // Initialize visualizer
+    setupVisualizer();
+    
+    // Show welcome message
+    addSystemMessage('Welcome to Sesame AI Voice Chat! Click "Start Conversation" to begin.');
+}
+
+// Initialize UI elements
+function initializeUIElements() {
+    // Store references to UI elements
+    elements.conversation = document.getElementById('conversation');
+    elements.streamButton = document.getElementById('streamButton');
+    elements.clearButton = document.getElementById('clearButton');
+    elements.thresholdSlider = document.getElementById('thresholdSlider');
+    elements.thresholdValue = document.getElementById('thresholdValue');
+    elements.visualizerCanvas = document.getElementById('audioVisualizer');
+    elements.visualizerLabel = document.getElementById('visualizerLabel');
+    elements.volumeLevel = document.getElementById('volumeLevel');
+    elements.statusDot = document.getElementById('statusDot');
+    elements.statusText = document.getElementById('statusText');
+    elements.speakerSelection = document.getElementById('speakerSelect'); // Changed to match HTML
+    elements.autoPlayResponses = document.getElementById('autoPlayResponses');
+    elements.showVisualizer = document.getElementById('showVisualizer');
+}
+
+// Setup Socket.IO connection
+function setupSocketConnection() {
+    state.socket = io(SERVER_URL);
+    
+    // Connection events
+    state.socket.on('connect', () => {
+        console.log('Connected to server');
+        updateConnectionStatus(true);
+    });
+    
+    state.socket.on('disconnect', () => {
+        console.log('Disconnected from server');
+        updateConnectionStatus(false);
+        
+        // Stop streaming if active
+        if (state.isStreaming) {
+            stopStreaming(false);
+        }
+    });
+    
+    state.socket.on('error', (data) => {
+        console.error('Socket error:', data.message);
+        addSystemMessage(`Error: ${data.message}`);
+    });
+    
+    // Register message handlers
+    state.socket.on('audio_response', handleAudioResponse);
+    state.socket.on('transcription', handleTranscription);
+    state.socket.on('context_updated', handleContextUpdate);
+    state.socket.on('streaming_status', handleStreamingStatus);
+}
+
+// Setup event listeners
+function setupEventListeners() {
+    // Stream button
+    elements.streamButton.addEventListener('click', toggleStreaming);
+    
+    // Clear button
+    elements.clearButton.addEventListener('click', clearConversation);
+    
+    // Threshold slider
+    elements.thresholdSlider.addEventListener('input', updateThreshold);
+    
+    // Speaker selection
+    elements.speakerSelection.addEventListener('change', () => {
+        state.currentSpeaker = parseInt(elements.speakerSelection.value, 10);
+    });
+    
+    // Visualizer toggle
+    elements.showVisualizer.addEventListener('change', toggleVisualizerVisibility);
+}
+
+// Setup audio visualizer
+function setupVisualizer() {
+    if (!elements.visualizerCanvas) return;
+    
+    canvasContext = elements.visualizerCanvas.getContext('2d');
+    
+    // Set canvas dimensions
+    elements.visualizerCanvas.width = elements.visualizerCanvas.offsetWidth;
+    elements.visualizerCanvas.height = elements.visualizerCanvas.offsetHeight;
+    
+    // Initialize the visualizer
+    drawVisualizer();
+}
+
+// Update connection status UI
+function updateConnectionStatus(isConnected) {
+    elements.statusDot.classList.toggle('active', isConnected);
+    elements.statusText.textContent = isConnected ? 'Connected' : 'Disconnected';
+}
+
+// Toggle streaming state
+function toggleStreaming() {
+    if (state.isStreaming) {
+        stopStreaming(true);
+    } else {
+        startStreaming();
+    }
+}
+
+// Start streaming audio to the server
+function startStreaming() {
+    if (state.isStreaming) return;
+    
+    // Request microphone access
+    navigator.mediaDevices.getUserMedia({ audio: true, video: false })
+        .then(stream => {
+            // Show processing state while setting up
+            elements.streamButton.innerHTML = '<i class="fas fa-spinner fa-spin"></i> Initializing...';
+            
+            // Create audio context
+            state.audioContext = new (window.AudioContext || window.webkitAudioContext)();
+            
+            // Create microphone source
+            state.microphone = state.audioContext.createMediaStreamSource(stream);
+            
+            // Create analyser for visualizer
+            state.analyser = state.audioContext.createAnalyser();
+            state.analyser.fftSize = 256;
+            visualizerBufferLength = state.analyser.frequencyBinCount;
+            visualizerDataArray = new Uint8Array(visualizerBufferLength);
+            
+            // Connect microphone to analyser
+            state.microphone.connect(state.analyser);
+            
+            // Create script processor for audio processing
+            const bufferSize = 4096;
+            state.streamProcessor = state.audioContext.createScriptProcessor(bufferSize, 1, 1);
+            
+            // Set up audio processing callback
+            state.streamProcessor.onaudioprocess = handleAudioProcess;
+            
+            // Connect the processors
+            state.analyser.connect(state.streamProcessor);
+            state.streamProcessor.connect(state.audioContext.destination);
+            
+            // Update UI
+            state.isStreaming = true;
+            elements.streamButton.innerHTML = '<i class="fas fa-microphone"></i> Listening...';
+            elements.streamButton.classList.add('recording');
+            
+            // Initialize energy window
+            state.energyWindow = [];
+            
+            // Start volume meter updates
+            state.volumeUpdateInterval = setInterval(updateVolumeMeter, 100);
+            
+            // Start visualizer if enabled
+            if (elements.showVisualizer.checked && !state.visualizerAnimationFrame) {
+                drawVisualizer();
+            }
+            
+            // Show starting message
+            addSystemMessage('Listening... Speak clearly into your microphone.');
+            
+            // Notify the server that we're starting
+            state.socket.emit('stream_audio', {
+                audio: '',
+                speaker: state.currentSpeaker
+            });
+        })
+        .catch(err => {
+            console.error('Error accessing microphone:', err);
+            addSystemMessage(`Error: ${err.message}. Please make sure your microphone is connected and you've granted permission.`);
+            elements.streamButton.innerHTML = '<i class="fas fa-microphone"></i> Start Conversation';
+        });
+}
+
+// Stop streaming audio
+function stopStreaming(notifyServer = true) {
+    if (!state.isStreaming) return;
+    
+    // Update UI first
+    elements.streamButton.innerHTML = '<i class="fas fa-microphone"></i> Start Conversation';
+    elements.streamButton.classList.remove('recording');
+    elements.streamButton.classList.remove('processing');
+    
+    // Stop volume meter updates
+    if (state.volumeUpdateInterval) {
+        clearInterval(state.volumeUpdateInterval);
+        state.volumeUpdateInterval = null;
+    }
+    
+    // Stop all audio processing
+    if (state.streamProcessor) {
+        state.streamProcessor.disconnect();
+        state.streamProcessor = null;
+    }
+    
+    if (state.analyser) {
+        state.analyser.disconnect();
+    }
+    
+    if (state.microphone) {
+        state.microphone.disconnect();
+    }
+    
+    // Close audio context
+    if (state.audioContext && state.audioContext.state !== 'closed') {
+        state.audioContext.close().catch(err => console.warn('Error closing audio context:', err));
+    }
+    
+    // Cleanup animation frames
+    if (state.visualizerAnimationFrame) {
+        cancelAnimationFrame(state.visualizerAnimationFrame);
+        state.visualizerAnimationFrame = null;
+    }
+    
+    // Reset state
+    state.isStreaming = false;
+    state.isSpeaking = false;
+    
+    // Notify the server
+    if (notifyServer && state.socket && state.socket.connected) {
+        state.socket.emit('stop_streaming', {
+            speaker: state.currentSpeaker
+        });
+    }
+    
+    // Show message
+    addSystemMessage('Conversation paused. Click "Start Conversation" to resume.');
+}
+
+// Handle audio processing
+function handleAudioProcess(event) {
+    const inputData = event.inputBuffer.getChannelData(0);
+    
+    // Log audio buffer statistics
+    console.log(`Audio buffer: length=${inputData.length}, sample rate=${state.audioContext.sampleRate}Hz`);
+    
+    // Calculate audio energy (volume level)
+    const energy = calculateAudioEnergy(inputData);
+    console.log(`Energy: ${energy.toFixed(6)}, threshold: ${state.silenceThreshold}`);
+    
+    // Update energy window for averaging
+    updateEnergyWindow(energy);
+    
+    // Calculate average energy
+    const avgEnergy = calculateAverageEnergy();
+    
+    // Determine if audio is silent
+    const isSilent = avgEnergy < state.silenceThreshold;
+    console.log(`Silent: ${isSilent ? 'Yes' : 'No'}, avg energy: ${avgEnergy.toFixed(6)}`);
+    
+    // Handle speech state based on silence
+    handleSpeechState(isSilent);
+    
+    // Only send audio chunk if we detect speech
+    if (!isSilent) {
+        // Create a resampled version at 24kHz for the server
+        // Most WebRTC audio is 48kHz, but we want 24kHz for the model
+        const resampledData = downsampleBuffer(inputData, state.audioContext.sampleRate, 24000);
+        console.log(`Resampled audio: ${state.audioContext.sampleRate}Hz → 24000Hz, new length: ${resampledData.length}`);
+        
+        // Send the audio chunk to the server
+        sendAudioChunk(resampledData, state.currentSpeaker);
+    }
+}
+
+// Cleanup audio resources when done
+function cleanupAudioResources() {
+    // Stop all audio processing
+    if (state.streamProcessor) {
+        state.streamProcessor.disconnect();
+        state.streamProcessor = null;
+    }
+    
+    if (state.analyser) {
+        state.analyser.disconnect();
+        state.analyser = null;
+    }
+    
+    if (state.microphone) {
+        state.microphone.disconnect();
+        state.microphone = null;
+    }
+    
+    // Close audio context
+    if (state.audioContext && state.audioContext.state !== 'closed') {
+        state.audioContext.close().catch(err => console.warn('Error closing audio context:', err));
+    }
+    
+    // Cancel all timers and animation frames
+    if (state.volumeUpdateInterval) {
+        clearInterval(state.volumeUpdateInterval);
+        state.volumeUpdateInterval = null;
+    }
+    
+    if (state.visualizerAnimationFrame) {
+        cancelAnimationFrame(state.visualizerAnimationFrame);
+        state.visualizerAnimationFrame = null;
+    }
+    
+    if (state.silenceTimer) {
+        clearTimeout(state.silenceTimer);
+        state.silenceTimer = null;
+    }
+}
+
+// Clear conversation history
+function clearConversation() {
+    if (elements.conversation) {
+        elements.conversation.innerHTML = '';
+        addSystemMessage('Conversation cleared.');
+        
+        // Notify server to clear context
+        if (state.socket && state.socket.connected) {
+            state.socket.emit('clear_context');
+        }
+    }
+}
+
+// Calculate audio energy (volume)
+function calculateAudioEnergy(buffer) {
+    let sum = 0;
+    for (let i = 0; i < buffer.length; i++) {
+        sum += buffer[i] * buffer[i];
+    }
+    return Math.sqrt(sum / buffer.length);
+}
+
+// Update energy window for averaging
+function updateEnergyWindow(energy) {
+    state.energyWindow.push(energy);
+    if (state.energyWindow.length > ENERGY_WINDOW_SIZE) {
+        state.energyWindow.shift();
+    }
+}
+
+// Calculate average energy from window
+function calculateAverageEnergy() {
+    if (state.energyWindow.length === 0) return 0;
+    
+    const sum = state.energyWindow.reduce((a, b) => a + b, 0);
+    return sum / state.energyWindow.length;
+}
+
+// Update the threshold from the slider
+function updateThreshold() {
+    state.silenceThreshold = parseFloat(elements.thresholdSlider.value);
+    elements.thresholdValue.textContent = state.silenceThreshold.toFixed(3);
+}
+
+// Update the volume meter display
+function updateVolumeMeter() {
+    if (!state.isStreaming || !state.energyWindow.length) return;
+    
+    const avgEnergy = calculateAverageEnergy();
+    
+    // Scale energy to percentage (0-100)
+    // Typically, energy values will be very small (e.g., 0.001 to 0.1)
+    // So we multiply by a factor to make it more visible
+    const scaleFactor = 1000;
+    const percentage = Math.min(100, Math.max(0, avgEnergy * scaleFactor));
+    
+    // Update volume meter width
+    elements.volumeLevel.style.width = `${percentage}%`;
+    
+    // Change color based on level
+    if (percentage > 70) {
+        elements.volumeLevel.style.backgroundColor = '#ff5252';
+    } else if (percentage > 30) {
+        elements.volumeLevel.style.backgroundColor = '#4CAF50';
+    } else {
+        elements.volumeLevel.style.backgroundColor = '#4c84ff';
+    }
+}
+
+// Handle speech/silence state transitions
+function handleSpeechState(isSilent) {
+    if (state.isSpeaking && isSilent) {
+        // Transition from speaking to silence
+        if (!state.silenceTimer) {
+            state.silenceTimer = setTimeout(() => {
+                // Only consider it a real silence after a certain duration
+                // This prevents detecting brief pauses as the end of speech
+                state.isSpeaking = false;
+                state.silenceTimer = null;
+            }, CLIENT_SILENCE_DURATION_MS);
+        }
+    } else if (state.silenceTimer && !isSilent) {
+        // User started speaking again, cancel the silence timer
+        clearTimeout(state.silenceTimer);
+        state.silenceTimer = null;
+    }
+    
+    // Update speaking state for non-silent audio
+    if (!isSilent) {
+        state.isSpeaking = true;
+    }
+}
+
+// Send audio chunk to server
+function sendAudioChunk(audioData, speaker) {
+    if (!state.socket || !state.socket.connected) {
+        console.warn('Socket not connected');
+        return;
+    }
+    
+    console.log(`Preparing audio chunk: length=${audioData.length}, speaker=${speaker}`);
+    
+    // Check for NaN or invalid values
+    let hasInvalidValues = false;
+    for (let i = 0; i < audioData.length; i++) {
+        if (isNaN(audioData[i]) || !isFinite(audioData[i])) {
+            hasInvalidValues = true;
+            console.warn(`Invalid audio value at index ${i}: ${audioData[i]}`);
+            break;
+        }
+    }
+    
+    if (hasInvalidValues) {
+        console.warn('Audio data contains invalid values. Creating silent audio.');
+        audioData = new Float32Array(audioData.length).fill(0);
+    }
+    
+    try {
+        // Create WAV blob
+        const wavData = createWavBlob(audioData, 24000);
+        console.log(`WAV blob created: ${wavData.size} bytes`);
+        
+        const reader = new FileReader();
+        
+        reader.onloadend = function() {
+            try {
+                // Get base64 data
+                const base64data = reader.result;
+                console.log(`Base64 data created: ${base64data.length} bytes`);
+                
+                // Send to server
+                state.socket.emit('stream_audio', {
+                    audio: base64data,
+                    speaker: speaker
+                });
+                console.log('Audio chunk sent to server');
+            } catch (err) {
+                console.error('Error preparing audio data:', err);
+            }
+        };
+        
+        reader.onerror = function() {
+            console.error('Error reading audio data as base64');
+        };
+        
+        reader.readAsDataURL(wavData);
+    } catch (err) {
+        console.error('Error creating WAV data:', err);
+    }
+}
+
+// Create WAV blob from audio data with improved error handling
+function createWavBlob(audioData, sampleRate) {
+    // Validate input
+    if (!audioData || audioData.length === 0) {
+        console.warn('Empty audio data provided to createWavBlob');
+        audioData = new Float32Array(1024).fill(0); // Create 1024 samples of silence
+    }
+    
+    // Function to convert Float32Array to Int16Array for WAV format
+    function floatTo16BitPCM(output, offset, input) {
+        for (let i = 0; i < input.length; i++, offset += 2) {
+            // Ensure values are in -1 to 1 range
+            const s = Math.max(-1, Math.min(1, input[i]));
+            // Convert to 16-bit PCM
+            output.setInt16(offset, s < 0 ? s * 0x8000 : s * 0x7FFF, true);
+        }
+    }
+    
+    // Create WAV header
+    function writeString(view, offset, string) {
+        for (let i = 0; i < string.length; i++) {
+            view.setUint8(offset + i, string.charCodeAt(i));
+        }
+    }
+    
+    try {
+        // Create WAV file with header - careful with buffer sizes
+        const buffer = new ArrayBuffer(44 + audioData.length * 2);
+        const view = new DataView(buffer);
+        
+        // RIFF identifier
+        writeString(view, 0, 'RIFF');
+        
+        // File length (will be filled later)
+        view.setUint32(4, 36 + audioData.length * 2, true);
+        
+        // WAVE identifier
+        writeString(view, 8, 'WAVE');
+        
+        // fmt chunk identifier
+        writeString(view, 12, 'fmt ');
+        
+        // fmt chunk length
+        view.setUint32(16, 16, true);
+        
+        // Sample format (1 is PCM)
+        view.setUint16(20, 1, true);
+        
+        // Mono channel
+        view.setUint16(22, 1, true);
+        
+        // Sample rate
+        view.setUint32(24, sampleRate, true);
+        
+        // Byte rate (sample rate * block align)
+        view.setUint32(28, sampleRate * 2, true);
+        
+        // Block align (channels * bytes per sample)
+        view.setUint16(32, 2, true);
+        
+        // Bits per sample
+        view.setUint16(34, 16, true);
+        
+        // data chunk identifier
+        writeString(view, 36, 'data');
+        
+        // data chunk length
+        view.setUint32(40, audioData.length * 2, true);
+        
+        // Write the PCM samples
+        floatTo16BitPCM(view, 44, audioData);
+        
+        // Create and return blob
+        return new Blob([view], { type: 'audio/wav' });
+    } catch (err) {
+        console.error('Error in createWavBlob:', err);
+        
+        // Create a minimal valid WAV file with silence as fallback
+        const fallbackSamples = new Float32Array(1024).fill(0);
+        const fallbackBuffer = new ArrayBuffer(44 + fallbackSamples.length * 2);
+        const fallbackView = new DataView(fallbackBuffer);
+        
+        writeString(fallbackView, 0, 'RIFF');
+        fallbackView.setUint32(4, 36 + fallbackSamples.length * 2, true);
+        writeString(fallbackView, 8, 'WAVE');
+        writeString(fallbackView, 12, 'fmt ');
+        fallbackView.setUint32(16, 16, true);
+        fallbackView.setUint16(20, 1, true);
+        fallbackView.setUint16(22, 1, true);
+        fallbackView.setUint32(24, sampleRate, true);
+        fallbackView.setUint32(28, sampleRate * 2, true);
+        fallbackView.setUint16(32, 2, true);
+        fallbackView.setUint16(34, 16, true);
+        writeString(fallbackView, 36, 'data');
+        fallbackView.setUint32(40, fallbackSamples.length * 2, true);
+        floatTo16BitPCM(fallbackView, 44, fallbackSamples);
+        
+        return new Blob([fallbackView], { type: 'audio/wav' });
+    }
+}
+
+// Draw audio visualizer
+function drawVisualizer() {
+    if (!canvasContext) {
+        return;
+    }
+    
+    state.visualizerAnimationFrame = requestAnimationFrame(drawVisualizer);
+    
+    // Skip drawing if visualizer is hidden
+    if (!elements.showVisualizer.checked) {
+        if (elements.visualizerCanvas.style.opacity !== '0') {
+            elements.visualizerCanvas.style.opacity = '0';
+        }
+        return;
+    } else if (elements.visualizerCanvas.style.opacity !== '1') {
+        elements.visualizerCanvas.style.opacity = '1';
+    }
+    
+    // Get frequency data if available
+    if (state.isStreaming && state.analyser) {
+        try {
+            state.analyser.getByteFrequencyData(visualizerDataArray);
+        } catch (e) {
+            console.warn('Error getting frequency data:', e);
+        }
+    } else {
+        // Fade out when not streaming
+        for (let i = 0; i < visualizerDataArray.length; i++) {
+            visualizerDataArray[i] = Math.max(0, visualizerDataArray[i] - 5);
+        }
+    }
+    
+    // Clear canvas
+    canvasContext.fillStyle = 'rgb(0, 0, 0)';
+    canvasContext.fillRect(0, 0, elements.visualizerCanvas.width, elements.visualizerCanvas.height);
+    
+    // Draw gradient bars
+    const width = elements.visualizerCanvas.width;
+    const height = elements.visualizerCanvas.height;
+    const barCount = Math.min(visualizerBufferLength, 64);
+    const barWidth = width / barCount - 1;
+    
+    for (let i = 0; i < barCount; i++) {
+        const index = Math.floor(i * visualizerBufferLength / barCount);
+        const value = visualizerDataArray[index];
+        
+        // Use logarithmic scale for better audio visualization
+        // This makes low values more visible while still maintaining full range
+        const logFactor = 20;
+        const scaledValue = Math.log(1 + (value / 255) * logFactor) / Math.log(1 + logFactor);
+        const barHeight = scaledValue * height;
+        
+        // Position bars
+        const x = i * (barWidth + 1);
+        const y = height - barHeight;
+        
+        // Create color gradient based on frequency and amplitude
+        const hue = i / barCount * 360; // Full color spectrum
+        const saturation = 80 + (value / 255 * 20); // Higher values more saturated
+        const lightness = 40 + (value / 255 * 20); // Dynamic brightness based on amplitude
+        
+        // Draw main bar
+        canvasContext.fillStyle = `hsl(${hue}, ${saturation}%, ${lightness}%)`;
+        canvasContext.fillRect(x, y, barWidth, barHeight);
+        
+        // Add reflection effect
+        if (barHeight > 5) {
+            const gradient = canvasContext.createLinearGradient(
+                x, y, 
+                x, y + barHeight * 0.5
+            );
+            gradient.addColorStop(0, `hsla(${hue}, ${saturation}%, ${lightness + 20}%, 0.4)`);
+            gradient.addColorStop(1, `hsla(${hue}, ${saturation}%, ${lightness}%, 0)`);
+            canvasContext.fillStyle = gradient;
+            canvasContext.fillRect(x, y, barWidth, barHeight * 0.5);
+            
+            // Add highlight on top of the bar for better 3D effect
+            canvasContext.fillStyle = `hsla(${hue}, ${saturation - 20}%, ${lightness + 30}%, 0.7)`;
+            canvasContext.fillRect(x, y, barWidth, 2);
+        }
+    }
+    
+    // Show/hide the label
+    elements.visualizerLabel.style.opacity = (state.isStreaming) ? '0' : '0.7';
+}
+
+// Toggle visualizer visibility
+function toggleVisualizerVisibility() {
+    const isVisible = elements.showVisualizer.checked;
+    elements.visualizerCanvas.style.opacity = isVisible ? '1' : '0';
+    
+    if (isVisible && state.isStreaming && !state.visualizerAnimationFrame) {
+        drawVisualizer();
+    }
+}
+
+// Handle audio response from server
+function handleAudioResponse(data) {
+    console.log('Received audio response');
+    
+    // Create message container
+    const messageElement = document.createElement('div');
+    messageElement.className = 'message ai';
+    
+    // Add text content if available
+    if (data.text) {
+        const textElement = document.createElement('p');
+        textElement.textContent = data.text;
+        messageElement.appendChild(textElement);
+    }
+    
+    // Create and configure audio element
+    const audioElement = document.createElement('audio');
+    audioElement.controls = true;
+    audioElement.className = 'audio-player';
+    
+    // Set audio source
+    const audioSource = document.createElement('source');
+    audioSource.src = data.audio;
+    audioSource.type = 'audio/wav';
+    
+    // Add fallback text
+    audioElement.textContent = 'Your browser does not support the audio element.';
+    
+    // Assemble audio element
+    audioElement.appendChild(audioSource);
+    messageElement.appendChild(audioElement);
+    
+    // Add timestamp
+    const timeElement = document.createElement('span');
+    timeElement.className = 'message-time';
+    timeElement.textContent = new Date().toLocaleTimeString();
+    messageElement.appendChild(timeElement);
+    
+    // Add to conversation
+    elements.conversation.appendChild(messageElement);
+    
+    // Auto-scroll to bottom
+    elements.conversation.scrollTop = elements.conversation.scrollHeight;
+    
+    // Auto-play if enabled
+    if (elements.autoPlayResponses.checked) {
+        audioElement.play()
+            .catch(err => {
+                console.warn('Auto-play failed:', err);
+                addSystemMessage('Auto-play failed. Please click play to hear the response.');
+            });
+    }
+    
+    // Re-enable stream button after processing is complete
+    if (state.isStreaming) {
+        elements.streamButton.innerHTML = '<i class="fas fa-microphone"></i> Listening...';
+        elements.streamButton.classList.add('recording');
+        elements.streamButton.classList.remove('processing');
+    }
+}
+
+// Handle transcription response from server
+function handleTranscription(data) {
+    console.log('Received transcription:', data.text);
+    
+    // Create message element
+    const messageElement = document.createElement('div');
+    messageElement.className = 'message user';
+    
+    // Add text content
+    const textElement = document.createElement('p');
+    textElement.textContent = data.text;
+    messageElement.appendChild(textElement);
+    
+    // Add timestamp
+    const timeElement = document.createElement('span');
+    timeElement.className = 'message-time';
+    timeElement.textContent = new Date().toLocaleTimeString();
+    messageElement.appendChild(timeElement);
+    
+    // Add to conversation
+    elements.conversation.appendChild(messageElement);
+    
+    // Auto-scroll to bottom
+    elements.conversation.scrollTop = elements.conversation.scrollHeight;
+}
+
+// Handle context update from server
+function handleContextUpdate(data) {
+    console.log('Context updated:', data.message);
+}
+
+// Handle streaming status updates from server
+function handleStreamingStatus(data) {
+    console.log('Streaming status:', data.status);
+    
+    if (data.status === 'stopped') {
+        // Reset UI if needed
+        if (state.isStreaming) {
+            stopStreaming(false); // Don't send to server since this came from server
+        }
+    }
+}
+
+// Add a system message to the conversation
+function addSystemMessage(message) {
+    const messageElement = document.createElement('div');
+    messageElement.className = 'message system';
+    messageElement.textContent = message;
+    elements.conversation.appendChild(messageElement);
+    
+    // Auto-scroll to bottom
+    elements.conversation.scrollTop = elements.conversation.scrollHeight;
+}
+
+// Downsample audio buffer to target sample rate
+function downsampleBuffer(buffer, originalSampleRate, targetSampleRate) {
+    if (originalSampleRate === targetSampleRate) {
+        return buffer;
+    }
+    
+    const ratio = originalSampleRate / targetSampleRate;
+    const newLength = Math.round(buffer.length / ratio);
+    const result = new Float32Array(newLength);
+    
+    for (let i = 0; i < newLength; i++) {
+        const pos = Math.round(i * ratio);
+        result[i] = buffer[pos];
+    }
+    
+    return result;
+}
+
+// Initialize the application when DOM is fully loaded
+document.addEventListener('DOMContentLoaded', initializeApp);
+

React/bun.lock

@@ -5,6 +5,7 @@
       "name": "my-app",
       "dependencies": {
         "@auth0/nextjs-auth0": "^4.3.0",
+        "mongoose": "^8.13.1",
         "next": "15.2.4",
         "react": "^19.1.0",
         "react-dom": "^19.1.0",
@@ -68,6 +69,8 @@
 
     "@img/sharp-win32-x64": ["@img/sharp-win32-x64@0.33.5", "", { "os": "win32", "cpu": "x64" }, "sha512-MpY/o8/8kj+EcnxwvrP4aTJSWw/aZ7JIGR4aBeZkZw5B7/Jn+tY9/VNwtcoGmdT7GfggGIU4kygOMSbYnOrAbg=="],
 
+    "@mongodb-js/saslprep": ["@mongodb-js/saslprep@1.2.0", "", { "dependencies": { "sparse-bitfield": "^3.0.3" } }, "sha512-+ywrb0AqkfaYuhHs6LxKWgqbh3I72EpEgESCw37o+9qPx9WTCkgDm2B+eMrwehGtHBWHFU4GXvnSCNiFhhausg=="],
+
     "@next/env": ["@next/env@15.2.4", "", {}, "sha512-+SFtMgoiYP3WoSswuNmxJOCwi06TdWE733D+WPjpXIe4LXGULwEaofiiAy6kbS0+XjM5xF5n3lKuBwN2SnqD9g=="],
 
     "@next/swc-darwin-arm64": ["@next/swc-darwin-arm64@15.2.4", "", { "os": "darwin", "cpu": "arm64" }, "sha512-1AnMfs655ipJEDC/FHkSr0r3lXBgpqKo4K1kiwfUf3iE68rDFXZ1TtHdMvf7D0hMItgDZ7Vuq3JgNMbt/+3bYw=="],
@@ -130,6 +133,10 @@
 
     "@types/react-dom": ["@types/react-dom@19.0.4", "", { "peerDependencies": { "@types/react": "^19.0.0" } }, "sha512-4fSQ8vWFkg+TGhePfUzVmat3eC14TXYSsiiDSLI0dVLsrm9gZFABjPy/Qu6TKgl1tq1Bu1yDsuQgY3A3DOjCcg=="],
 
+    "@types/webidl-conversions": ["@types/webidl-conversions@7.0.3", "", {}, "sha512-CiJJvcRtIgzadHCYXw7dqEnMNRjhGZlYK05Mj9OyktqV8uVT8fD2BFOB7S1uwBE3Kj2Z+4UyPmFw/Ixgw/LAlA=="],
+
+    "@types/whatwg-url": ["@types/whatwg-url@11.0.5", "", { "dependencies": { "@types/webidl-conversions": "*" } }, "sha512-coYR071JRaHa+xoEvvYqvnIHaVqaYrLPbsufM9BF63HkwI5Lgmy2QR8Q5K/lYDYo5AK82wOvSOS0UsLTpTG7uQ=="],
+
     "@zag-js/accordion": ["@zag-js/accordion@1.7.0", "", { "dependencies": { "@zag-js/anatomy": "1.7.0", "@zag-js/core": "1.7.0", "@zag-js/dom-query": "1.7.0", "@zag-js/types": "1.7.0", "@zag-js/utils": "1.7.0" } }, "sha512-LNJOjLTW2KwrToXBrXIbNIAiISA94n0AdWp14H8RrskdokywmEGiC0GgWTGEJ7DNA6TGP6Ae5o9rJ4fHSmCsDQ=="],
 
     "@zag-js/anatomy": ["@zag-js/anatomy@1.7.0", "", {}, "sha512-fkRgH6vPCwykmRdV38uAJeTtJc8tayAnURfoovHAtB9bK0goagPbpdcYTNyGn8msul0h+KBloOtnw4obvX0nPw=="],
@@ -182,6 +189,8 @@
 
     "@zag-js/utils": ["@zag-js/utils@1.7.0", "", {}, "sha512-yIxvH5V27a1WuLgCxHX7qpdtFo8vTJaZLafBpSNfVYG4B8FaxTE+P7JAcpmAzs3UyXura/WfAY2eVWWVBpk9ZA=="],
 
+    "bson": ["bson@6.10.3", "", {}, "sha512-MTxGsqgYTwfshYWTRdmZRC+M7FnG1b4y7RO7p2k3X24Wq0yv1m77Wsj0BzlPzd/IowgESfsruQCUToa7vbOpPQ=="],
+
     "busboy": ["busboy@1.6.0", "", { "dependencies": { "streamsearch": "^1.1.0" } }, "sha512-8SFQbg/0hQ9xy3UNTB0YEnsNBbWfhf7RtnzpL7TkBiTBRfrQ9Fxcnz7VJsleJpyp6rVLvXiuORqjlHi5q+PYuA=="],
 
     "caniuse-lite": ["caniuse-lite@1.0.30001707", "", {}, "sha512-3qtRjw/HQSMlDWf+X79N206fepf4SOOU6SQLMaq/0KkZLmSjPxAkBOQQ+FxbHKfHmYLZFfdWsO3KA90ceHPSnw=="],
@@ -198,6 +207,8 @@
 
     "csstype": ["csstype@3.1.3", "", {}, "sha512-M1uQkMl8rQK/szD0LNhtqxIPLpimGm8sOBwU7lLnCpSbTyY3yeU1Vc7l4KT5zT4s/yOxHH5O7tIuuLOCnLADRw=="],
 
+    "debug": ["debug@4.4.0", "", { "dependencies": { "ms": "^2.1.3" } }, "sha512-6WTZ/IxCY/T6BALoZHaE4ctp9xm+Z5kY/pzYaCHRFeyVhojxlrm+46y68HA6hr0TcwEssoxNiDEUJQjfPZ/RYA=="],
+
     "dequal": ["dequal@2.0.3", "", {}, "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA=="],
 
     "detect-libc": ["detect-libc@2.0.3", "", {}, "sha512-bwy0MGW55bG41VqxxypOsdSdGqLwXPI/focwgTYCFMbdUiBAxLg9CFzG08sz2aqzknwiX7Hkl0bQENjg8iLByw=="],
@@ -212,6 +223,8 @@
 
     "jose": ["jose@5.10.0", "", {}, "sha512-s+3Al/p9g32Iq+oqXxkW//7jk2Vig6FF1CFqzVXoTUXt2qz89YWbL+OwS17NFYEvxC35n0FKeGO2LGYSxeM2Gg=="],
 
+    "kareem": ["kareem@2.6.3", "", {}, "sha512-C3iHfuGUXK2u8/ipq9LfjFfXFxAZMQJJq7vLS45r3D9Y2xQ/m4S8zaR4zMLFWh9AsNPXmcFfUDhTEO8UIC/V6Q=="],
+
     "lightningcss": ["lightningcss@1.29.2", "", { "dependencies": { "detect-libc": "^2.0.3" }, "optionalDependencies": { "lightningcss-darwin-arm64": "1.29.2", "lightningcss-darwin-x64": "1.29.2", "lightningcss-freebsd-x64": "1.29.2", "lightningcss-linux-arm-gnueabihf": "1.29.2", "lightningcss-linux-arm64-gnu": "1.29.2", "lightningcss-linux-arm64-musl": "1.29.2", "lightningcss-linux-x64-gnu": "1.29.2", "lightningcss-linux-x64-musl": "1.29.2", "lightningcss-win32-arm64-msvc": "1.29.2", "lightningcss-win32-x64-msvc": "1.29.2" } }, "sha512-6b6gd/RUXKaw5keVdSEtqFVdzWnU5jMxTUjA2bVcMNPLwSQ08Sv/UodBVtETLCn7k4S1Ibxwh7k68IwLZPgKaA=="],
 
     "lightningcss-darwin-arm64": ["lightningcss-darwin-arm64@1.29.2", "", { "os": "darwin", "cpu": "arm64" }, "sha512-cK/eMabSViKn/PG8U/a7aCorpeKLMlK0bQeNHmdb7qUnBkNPnL+oV5DjJUo0kqWsJUapZsM4jCfYItbqBDvlcA=="],
@@ -234,6 +247,20 @@
 
     "lightningcss-win32-x64-msvc": ["lightningcss-win32-x64-msvc@1.29.2", "", { "os": "win32", "cpu": "x64" }, "sha512-EdIUW3B2vLuHmv7urfzMI/h2fmlnOQBk1xlsDxkN1tCWKjNFjfLhGxYk8C8mzpSfr+A6jFFIi8fU6LbQGsRWjA=="],
 
+    "memory-pager": ["memory-pager@1.5.0", "", {}, "sha512-ZS4Bp4r/Zoeq6+NLJpP+0Zzm0pR8whtGPf1XExKLJBAczGMnSi3It14OiNCStjQjM6NU1okjQGSxgEZN8eBYKg=="],
+
+    "mongodb": ["mongodb@6.15.0", "", { "dependencies": { "@mongodb-js/saslprep": "^1.1.9", "bson": "^6.10.3", "mongodb-connection-string-url": "^3.0.0" }, "peerDependencies": { "@aws-sdk/credential-providers": "^3.188.0", "@mongodb-js/zstd": "^1.1.0 || ^2.0.0", "gcp-metadata": "^5.2.0", "kerberos": "^2.0.1", "mongodb-client-encryption": ">=6.0.0 <7", "snappy": "^7.2.2", "socks": "^2.7.1" }, "optionalPeers": ["@aws-sdk/credential-providers", "@mongodb-js/zstd", "gcp-metadata", "kerberos", "mongodb-client-encryption", "snappy", "socks"] }, "sha512-ifBhQ0rRzHDzqp9jAQP6OwHSH7dbYIQjD3SbJs9YYk9AikKEettW/9s/tbSFDTpXcRbF+u1aLrhHxDFaYtZpFQ=="],
+
+    "mongodb-connection-string-url": ["mongodb-connection-string-url@3.0.2", "", { "dependencies": { "@types/whatwg-url": "^11.0.2", "whatwg-url": "^14.1.0 || ^13.0.0" } }, "sha512-rMO7CGo/9BFwyZABcKAWL8UJwH/Kc2x0g72uhDWzG48URRax5TCIcJ7Rc3RZqffZzO/Gwff/jyKwCU9TN8gehA=="],
+
+    "mongoose": ["mongoose@8.13.1", "", { "dependencies": { "bson": "^6.10.3", "kareem": "2.6.3", "mongodb": "~6.15.0", "mpath": "0.9.0", "mquery": "5.0.0", "ms": "2.1.3", "sift": "17.1.3" } }, "sha512-sRqlXI+6jhr9/KicCOjet1VVPONFsOxTrh14tfueX5y3GJ2ihswc5ewUUojuwdSS/5koGXLIPmGivDSApVXflA=="],
+
+    "mpath": ["mpath@0.9.0", "", {}, "sha512-ikJRQTk8hw5DEoFVxHG1Gn9T/xcjtdnOKIU1JTmGjZZlg9LST2mBLmcX3/ICIbgJydT2GOc15RnNy5mHmzfSew=="],
+
+    "mquery": ["mquery@5.0.0", "", { "dependencies": { "debug": "4.x" } }, "sha512-iQMncpmEK8R8ncT8HJGsGc9Dsp8xcgYMVSbs5jgnm1lFHTZqMJTUWTDx1LBO8+mK3tPNZWFLBghQEIOULSTHZg=="],
+
+    "ms": ["ms@2.1.3", "", {}, "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA=="],
+
     "nanoid": ["nanoid@3.3.11", "", { "bin": { "nanoid": "bin/nanoid.cjs" } }, "sha512-N8SpfPUnUp1bK+PMYW8qSWdl9U+wwNWI4QKxOYDy9JAro3WMX7p2OeVRF9v+347pnakNevPmiHhNmZ2HbFA76w=="],
 
     "next": ["next@15.2.4", "", { "dependencies": { "@next/env": "15.2.4", "@swc/counter": "0.1.3", "@swc/helpers": "0.5.15", "busboy": "1.6.0", "caniuse-lite": "^1.0.30001579", "postcss": "8.4.31", "styled-jsx": "5.1.6" }, "optionalDependencies": { "@next/swc-darwin-arm64": "15.2.4", "@next/swc-darwin-x64": "15.2.4", "@next/swc-linux-arm64-gnu": "15.2.4", "@next/swc-linux-arm64-musl": "15.2.4", "@next/swc-linux-x64-gnu": "15.2.4", "@next/swc-linux-x64-musl": "15.2.4", "@next/swc-win32-arm64-msvc": "15.2.4", "@next/swc-win32-x64-msvc": "15.2.4", "sharp": "^0.33.5" }, "peerDependencies": { "@opentelemetry/api": "^1.1.0", "@playwright/test": "^1.41.2", "babel-plugin-react-compiler": "*", "react": "^18.2.0 || 19.0.0-rc-de68d2f4-20241204 || ^19.0.0", "react-dom": "^18.2.0 || 19.0.0-rc-de68d2f4-20241204 || ^19.0.0", "sass": "^1.3.0" }, "optionalPeers": ["@opentelemetry/api", "@playwright/test", "babel-plugin-react-compiler", "sass"], "bin": { "next": "dist/bin/next" } }, "sha512-VwL+LAaPSxEkd3lU2xWbgEOtrM8oedmyhBqaVNmgKB+GvZlCy9rgaEc+y2on0wv+l0oSFqLtYD6dcC1eAedUaQ=="],
@@ -246,6 +273,8 @@
 
     "proxy-compare": ["proxy-compare@3.0.1", "", {}, "sha512-V9plBAt3qjMlS1+nC8771KNf6oJ12gExvaxnNzN/9yVRLdTv/lc+oJlnSzrdYDAvBfTStPCoiaCOTmTs0adv7Q=="],
 
+    "punycode": ["punycode@2.3.1", "", {}, "sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg=="],
+
     "react": ["react@19.1.0", "", {}, "sha512-FS+XFBNvn3GTAWq26joslQgWNoFu08F4kl0J4CgdNKADkdSGXQyTCnKteIAJy96Br6YbpEU1LSzV5dYtjMkMDg=="],
 
     "react-dom": ["react-dom@19.1.0", "", { "dependencies": { "scheduler": "^0.26.0" }, "peerDependencies": { "react": "^19.1.0" } }, "sha512-Xs1hdnE+DyKgeHJeJznQmYMIBG3TKIHJJT95Q58nHLSrElKlGQqDTR2HQ9fx5CN/Gk6Vh/kupBTDLU11/nDk/g=="],
@@ -256,10 +285,14 @@
 
     "sharp": ["sharp@0.33.5", "", { "dependencies": { "color": "^4.2.3", "detect-libc": "^2.0.3", "semver": "^7.6.3" }, "optionalDependencies": { "@img/sharp-darwin-arm64": "0.33.5", "@img/sharp-darwin-x64": "0.33.5", "@img/sharp-libvips-darwin-arm64": "1.0.4", "@img/sharp-libvips-darwin-x64": "1.0.4", "@img/sharp-libvips-linux-arm": "1.0.5", "@img/sharp-libvips-linux-arm64": "1.0.4", "@img/sharp-libvips-linux-s390x": "1.0.4", "@img/sharp-libvips-linux-x64": "1.0.4", "@img/sharp-libvips-linuxmusl-arm64": "1.0.4", "@img/sharp-libvips-linuxmusl-x64": "1.0.4", "@img/sharp-linux-arm": "0.33.5", "@img/sharp-linux-arm64": "0.33.5", "@img/sharp-linux-s390x": "0.33.5", "@img/sharp-linux-x64": "0.33.5", "@img/sharp-linuxmusl-arm64": "0.33.5", "@img/sharp-linuxmusl-x64": "0.33.5", "@img/sharp-wasm32": "0.33.5", "@img/sharp-win32-ia32": "0.33.5", "@img/sharp-win32-x64": "0.33.5" } }, "sha512-haPVm1EkS9pgvHrQ/F3Xy+hgcuMV0Wm9vfIBSiwZ05k+xgb0PkBQpGsAA/oWdDobNaZTH5ppvHtzCFbnSEwHVw=="],
 
+    "sift": ["sift@17.1.3", "", {}, "sha512-Rtlj66/b0ICeFzYTuNvX/EF1igRbbnGSvEyT79McoZa/DeGhMyC5pWKOEsZKnpkqtSeovd5FL/bjHWC3CIIvCQ=="],
+
     "simple-swizzle": ["simple-swizzle@0.2.2", "", { "dependencies": { "is-arrayish": "^0.3.1" } }, "sha512-JA//kQgZtbuY83m+xT+tXJkmJncGMTFT+C+g2h2R9uxkYIrE2yy9sgmcLhCnw57/WSD+Eh3J97FPEDFnbXnDUg=="],
 
     "source-map-js": ["source-map-js@1.2.1", "", {}, "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA=="],
 
+    "sparse-bitfield": ["sparse-bitfield@3.0.3", "", { "dependencies": { "memory-pager": "^1.0.2" } }, "sha512-kvzhi7vqKTfkh0PZU+2D2PIllw2ymqJKujUcyPMd9Y75Nv4nPbGJZXNhxsgdQab2BmlDct1YnfQCguEvHr7VsQ=="],
+
     "streamsearch": ["streamsearch@1.1.0", "", {}, "sha512-Mcc5wHehp9aXz1ax6bZUyY5afg9u2rv5cqQI3mRrYkGC8rW2hM02jWuwjtL++LS5qinSyhj2QfLyNsuc+VsExg=="],
 
     "styled-jsx": ["styled-jsx@5.1.6", "", { "dependencies": { "client-only": "0.0.1" }, "peerDependencies": { "react": ">= 16.8.0 || 17.x.x || ^18.0.0-0 || ^19.0.0-0" } }, "sha512-qSVyDTeMotdvQYoHWLNGwRFJHC+i+ZvdBRYosOFgC+Wg1vx4frN2/RG/NA7SYqqvKNLf39P2LSRA2pu6n0XYZA=="],
@@ -270,6 +303,8 @@
 
     "tapable": ["tapable@2.2.1", "", {}, "sha512-GNzQvQTOIP6RyTfE2Qxb8ZVlNmw0n88vp1szwWRimP02mnTsx3Wtn5qRdqY9w2XduFNUgvOwhNnQsjwCp+kqaQ=="],
 
+    "tr46": ["tr46@5.1.0", "", { "dependencies": { "punycode": "^2.3.1" } }, "sha512-IUWnUK7ADYR5Sl1fZlO1INDUhVhatWl7BtJWsIhwJ0UAK7ilzzIa8uIqOO/aYVWHZPJkKbEL+362wrzoeRF7bw=="],
+
     "tslib": ["tslib@2.8.1", "", {}, "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w=="],
 
     "typescript": ["typescript@5.8.2", "", { "bin": { "tsc": "bin/tsc", "tsserver": "bin/tsserver" } }, "sha512-aJn6wq13/afZp/jT9QZmwEjDqqvSGp1VT5GVg+f/t6/oVyrgXM6BY1h9BRh/O5p3PlUPAe+WuiEZOmb/49RqoQ=="],
@@ -278,6 +313,10 @@
 
     "use-sync-external-store": ["use-sync-external-store@1.5.0", "", { "peerDependencies": { "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0" } }, "sha512-Rb46I4cGGVBmjamjphe8L/UnvJD+uPPtTkNvX5mZgqdbavhI4EbgIWJiIHXJ8bc/i9EQGPRh4DwEURJ552Do0A=="],
 
+    "webidl-conversions": ["webidl-conversions@7.0.0", "", {}, "sha512-VwddBukDzu71offAQR975unBIGqfKZpM+8ZX6ySk8nYhVoo5CYaZyzt3YBvYtRtO+aoGlqxPg/B87NGVZ/fu6g=="],
+
+    "whatwg-url": ["whatwg-url@14.2.0", "", { "dependencies": { "tr46": "^5.1.0", "webidl-conversions": "^7.0.0" } }, "sha512-De72GdQZzNTUBBChsXueQUnPKDkg/5A5zp7pFDuQAj5UFoENpiACU0wlCvzpAGnTkj++ihpKwKyYewn/XNUbKw=="],
+
     "next/postcss": ["postcss@8.4.31", "", { "dependencies": { "nanoid": "^3.3.6", "picocolors": "^1.0.0", "source-map-js": "^1.0.2" } }, "sha512-PS08Iboia9mts/2ygV3eLpY5ghnUcfLV/EXTOW1E2qYxJKGGBUtNjN76FYHnMs36RmARn41bC0AZmn+rR0OVpQ=="],
   }
 }

React/package.json

@@ -10,6 +10,7 @@
   },
   "dependencies": {
     "@auth0/nextjs-auth0": "^4.3.0",
+    "mongoose": "^8.13.1",
     "next": "15.2.4",
     "react": "^19.1.0",
     "react-dom": "^19.1.0"

React/src/app/page.tsx

@@ -40,14 +40,29 @@ export default async function Home() {
 					type="submit">Set codeword</button>
 				</form>
 				{/* form for adding contacts */}
-				<form className="flex flex-col gap-[32px] row-start-2 items-center sm:items-start" onSubmit={(e) => e.preventDefault()}>
+				<form className="space-y-5 flex flex-col gap-[32px] row-start-2 items-center sm:items-start" onSubmit={(e) => e.preventDefault()}>
 					<input
 						type="text"
 						value={contacts}
 						onChange={(e) => setContacts(e.target.value.split(","))}
-						placeholder="contacts (comma separated)"
+						placeholder="Write down an emergency contact"
 						className="border border-gray-300 rounded-md p-2"
 					/>
+					<input
+						type="text"
+						value={contacts}
+						onChange={(e) => setContacts(e.target.value.split(","))}
+						placeholder="Write down an emergency contact"
+						className="border border-gray-300 rounded-md p-2"
+					/>
+					<input
+						type="text"
+						value={contacts}
+						onChange={(e) => setContacts(e.target.value.split(","))}
+						placeholder="Write down an emergency contact"
+						className="border border-gray-300 rounded-md p-2"
+					/>
+					<button type="button">Add</button>
 					<button className="bg-slate-500 text-yellow-300 text-stretch-50% font-lg rounded-md p-2" type="submit">Set contacts</button>
 				</form>
 			</div>
@@ -76,14 +91,45 @@ export default async function Home() {
 						type="submit">Set codeword</button>
 					</form>
 					{/* form for adding contacts */}
-					<form className="flex flex-col gap-[32px] row-start-2 items-center sm:items-start" onSubmit={(e) => e.preventDefault()}>
+					<form id="Contacts" className="space-y-5 flex flex-col gap-[32px] row-start-2 items-center sm:items-start" onSubmit={(e) => e.preventDefault()}>
-						<input
+					<input
-							type="text"
+						type="text"
-							value={contacts}
+						value={contacts}
-							onChange={(e) => setContacts(e.target.value.split(","))}
+						onChange={(e) => setContacts(e.target.value.split(","))}
-							placeholder="contacts (comma separated)"
+						placeholder="Write down an emergency contact"
-							className="border border-gray-300 rounded-md p-2"
+						className="border border-gray-300 rounded-md p-2"
-						/>
+					/>
+					<input
+						type="text"
+						value={contacts}
+						onChange={(e) => setContacts(e.target.value.split(","))}
+						placeholder="Write down an emergency contact"
+						className="border border-gray-300 rounded-md p-2"
+					/>
+					<input
+						type="text"
+						value={contacts}
+						onChange={(e) => setContacts(e.target.value.split(","))}
+						placeholder="Write down an emergency contact"
+						className="border border-gray-300 rounded-md p-2"
+					/>
+					<input
+						type="text"
+						value={contacts}
+						onChange={(e) => setContacts(e.target.value.split(","))}
+						placeholder="Write down an emergency contact"
+						className="text-input border border-gray-300 rounded-md p-2"
+					/>
+					<button onClick={() => {
+						alert("Adding contact...");
+						let elem = document.getElementsByClassName("text-input")[0] as HTMLElement;
+						console.log("Element:", elem);
+						let d = elem.cloneNode(true) as HTMLElement;
+						document.getElementById("Contacts")?.appendChild(d);
+					}}
+					className="bg-emerald-500 text-fuchsia-300"
+					type="button">Add</button>
+					
 						<button className="bg-slate-500 text-yellow-300 text-stretch-50% font-lg rounded-md p-2" type="submit">Set contacts</button>
 					</form>