add: urdu, llm

.gitignore

@@ -0,0 +1,3 @@
+.idea
+__pycache__
+.env

app.py

@@ -22,9 +22,12 @@ from google.genai.types import (
     PrebuiltVoiceConfig,
     SpeechConfig,
     VoiceConfig,
+    Content,
+    Part
 )
 from gradio.utils import get_space
 from pydantic import BaseModel
+
 current_dir = pathlib.Path(__file__).parent
 load_dotenv()
 api_key = os.getenv("GEMINI_API_KEY")
@@ -32,7 +35,6 @@ if not api_key:
     raise ValueError("GEMINI_API_KEY environment variable is not set")
 
 
-
 def encode_audio(data: np.ndarray) -> str:
     """Encode Audio data to send to the server"""
     return base64.b64encode(data.tobytes()).decode("UTF-8")
@@ -42,10 +44,10 @@ class GeminiHandler(AsyncStreamHandler):
     """Handler for the Gemini API"""
 
     def __init__(
-        self,
-        expected_layout: Literal["mono"] = "mono",
-        output_sample_rate: int = 24000,
-        output_frame_size: int = 480,
+            self,
+            expected_layout: Literal["mono"] = "mono",
+            output_sample_rate: int = 24000,
+            output_frame_size: int = 480,
     ) -> None:
         super().__init__(
             expected_layout,
@@ -82,16 +84,40 @@ class GeminiHandler(AsyncStreamHandler):
             speech_config=SpeechConfig(
                 voice_config=VoiceConfig(
                     prebuilt_voice_config=PrebuiltVoiceConfig(
-                        voice_name=voice_name,  # Now voice_name is a string
+                        voice_name=voice_name,
                     )
                 )
+            ),
+
+            system_instruction=Content(
+    parts=[Part(
+        text="""You are an AI calling assistant for Ishwor Subedi, an AI/ML freelancer. When speaking with clients:
+
+2. For professional inquiries, highlight these key skills concisely:
+   - 2+ years in machine learning and AI
+   - Computer Vision expertise
+   - NLP capabilities
+   - Software and mobile app development
+   - Upwork freelancer with proven track record
+3. For generic questions:
+   - Provide brief, direct answers (1-2 sentences)
+   - Avoid lengthy explanations
+   - Always connect responses back to Ishwor's services when possible
+4. Keep website reference simple: "Visit ishwor-subedi.com.np for portfolio details"
+5. Speak in Hindi throughout
+6. For unrelated topics: "Please contact Ishwor directly for assistance with this"
+
+Maintain professional tone while keeping all responses concise and focused.
+""")],
+                role="user"
+
             )
         )
         async with client.aio.live.connect(
-            model="gemini-2.0-flash-exp", config=config
+                model="gemini-2.0-flash-exp", config=config
         ) as session:
             async for audio in session.start_stream(
-                stream=self.stream(), mime_type="audio/pcm"
+                    stream=self.stream(), mime_type="audio/pcm"
             ):
                 if audio.data:
                     array = np.frombuffer(audio.data, dtype=np.int16)
@@ -123,10 +149,9 @@ stream = Stream(
     mode="send-receive",
     handler=GeminiHandler(),
     rtc_configuration=get_twilio_turn_credentials() if get_space() else None,
-    concurrency_limit=5 if get_space() else None,
+    concurrency_limit=2,
     time_limit=90 if get_space() else None,
     additional_inputs=[
-        
         gr.Dropdown(
             label="Voice",
             choices=[
@@ -147,7 +172,6 @@ class InputData(BaseModel):
     voice_name: str
 
 
-
 app = FastAPI()
 
 stream.mount(app)
@@ -155,7 +179,7 @@ stream.mount(app)
 
 @app.post("/input_hook")
 async def _(body: InputData):
-    stream.set_input(body.webrtc_id,  body.voice_name)
+    stream.set_input(body.webrtc_id, body.voice_name)
     return {"status": "ok"}
 
 
@@ -181,4 +205,4 @@ if __name__ == "__main__":
     elif mode == "PHONE":
         stream.fastphone(host="localhost", port=7860)
     else:
-        uvicorn.run(app, host="localhost", port=7860)
+        uvicorn.run(app, host="localhost", port=7860)

index.html

@@ -298,6 +298,108 @@
         .footer a:hover {
             text-decoration: underline;
         }
+
+        .subtitle {
+            font-size: 1.25rem;
+            color: var(--color-text-secondary);
+            margin-bottom: 1rem;
+        }
+
+        .language-badge {
+            display: inline-block;
+            padding: 0.5rem 1rem;
+            background: var(--gradient);
+            border-radius: 2rem;
+            font-size: 1rem;
+            margin-top: 1rem;
+            box-shadow: var(--box-shadow);
+        }
+
+        .feature-badges {
+            display: flex;
+            justify-content: center;
+            gap: 1rem;
+            margin-bottom: 2rem;
+            flex-wrap: wrap;
+        }
+
+        .badge {
+            display: flex;
+            align-items: center;
+            gap: 0.5rem;
+            padding: 0.75rem 1.25rem;
+            background-color: var(--color-surface-light);
+            border-radius: 1rem;
+            font-size: 0.875rem;
+            border: 1px solid rgba(255, 255, 255, 0.05);
+        }
+
+        .badge i {
+            color: var(--color-primary-light);
+        }
+
+        .developer-info {
+            display: flex;
+            align-items: center;
+            gap: 1rem;
+            margin-bottom: 1rem;
+        }
+
+        .developer-avatar {
+            width: 48px;
+            height: 48px;
+            border-radius: 50%;
+            border: 2px solid var(--color-primary);
+        }
+
+        .developer-details {
+            text-align: left;
+        }
+
+        .developer-title {
+            font-size: 0.75rem;
+            color: var(--color-primary-light);
+            margin-top: 0.25rem;
+        }
+
+        .social-links {
+            display: flex;
+            gap: 1rem;
+            margin-top: 1rem;
+        }
+
+        .social-links a {
+            color: var(--color-text-secondary);
+            font-size: 1.25rem;
+            transition: color 0.2s ease;
+        }
+
+        .social-links a:hover {
+            color: var(--color-primary-light);
+        }
+
+        .footer {
+            background-color: var(--color-surface);
+            padding: 1.5rem;
+            border-radius: 1rem;
+            margin-top: 2rem;
+            box-shadow: var(--box-shadow);
+            border: 1px solid rgba(255, 255, 255, 0.05);
+            max-width: 600px;
+            width: 90%;
+        }
+
+        @media (max-width: 768px) {
+            .feature-badges {
+                flex-direction: column;
+                align-items: center;
+            }
+            
+            .badge {
+                width: 100%;
+                justify-content: center;
+            }
+        }
     </style>
 </head>
 
@@ -305,11 +407,30 @@
     <div id="error-toast" class="toast"></div>
     
     <div class="header">
-        <h1>Voice Assistant</h1>
-        <p>Real-time voice conversations with advanced AI technology</p>
+        <h1>Urdu AI Assistant</h1>
+        <p class="subtitle">Real-time Urdu conversations powered by advanced AI technology</p>
+        <div class="language-badge">
+            <i class="fas fa-language"></i>
+            हिंदी / اردو
+        </div>
     </div>
     
     <div class="container">
+        <div class="feature-badges">
+            <div class="badge">
+                <i class="fas fa-microphone-alt"></i>
+                Real-time Voice
+            </div>
+            <div class="badge">
+                <i class="fas fa-brain"></i>
+                AI Powered
+            </div>
+            <div class="badge">
+                <i class="fas fa-language"></i>
+                Urdu Support
+            </div>
+        </div>
+
         <div class="controls">
             <!-- <div class="input-group">
                 <label for="api-key"><i class="fas fa-key"></i> API Key</label>
@@ -337,7 +458,18 @@
     </div>
 
     <div class="footer">
-        <p>Powered by advanced AI technology</p>
+        <div class="developer-info">
+            <img src="https://media.licdn.com/dms/image/v2/D4D03AQFccVUQVirANg/profile-displayphoto-shrink_800_800/profile-displayphoto-shrink_800_800/0/1724430812003?e=1746662400&v=beta&t=w7dT8KntFGMuWX8sIb6J53xSsiV-M_MXLYBnBNVCARg" alt="Ishwor Subedi" class="developer-avatar">
+            <div class="developer-details">
+                <p>Developed by <a href="https://ishwor-subedi.com.np" target="_blank">Ishwor Subedi</a></p>
+                <p class="developer-title">AI/ML Engineer & Freelancer</p>
+            </div>
+        </div>
+        <div class="social-links">
+            <a href="https://github.com/ishworrsubedii" target="_blank"><i class="fab fa-github"></i></a>
+            <a href="https://www.linkedin.com/in/ishworrsubedii/" target="_blank"><i class="fab fa-linkedin"></i></a>
+            <a href="https://www.upwork.com/freelancers/~01a5bd20f3bdbf5bde" target="_blank"><i class="fab fa-upwork"></i></a>
+        </div>
     </div>
 
     <audio id="audio-output"></audio>

index2.html

@@ -1,496 +0,0 @@
-<!DOCTYPE html>
-<html lang="en">
-
-<head>
-    <meta charset="UTF-8">
-    <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>Gemini Voice Chat</title>
-    <style>
-        :root {
-            --color-accent: #6366f1;
-            --color-background: #0f172a;
-            --color-surface: #1e293b;
-            --color-text: #e2e8f0;
-            --boxSize: 8px;
-            --gutter: 4px;
-        }
-
-        body {
-            margin: 0;
-            padding: 0;
-            background-color: var(--color-background);
-            color: var(--color-text);
-            font-family: system-ui, -apple-system, sans-serif;
-            min-height: 100vh;
-            display: flex;
-            flex-direction: column;
-            align-items: center;
-            justify-content: center;
-        }
-
-        .container {
-            width: 90%;
-            max-width: 800px;
-            background-color: var(--color-surface);
-            padding: 2rem;
-            border-radius: 1rem;
-            box-shadow: 0 25px 50px -12px rgba(0, 0, 0, 0.25);
-        }
-
-        .wave-container {
-            position: relative;
-            display: flex;
-            min-height: 100px;
-            max-height: 128px;
-            justify-content: center;
-            align-items: center;
-            margin: 2rem 0;
-        }
-
-        .box-container {
-            display: flex;
-            justify-content: space-between;
-            height: 64px;
-            width: 100%;
-        }
-
-        .box {
-            height: 100%;
-            width: var(--boxSize);
-            background: var(--color-accent);
-            border-radius: 8px;
-            transition: transform 0.05s ease;
-        }
-
-        .controls {
-            display: grid;
-            gap: 1rem;
-            margin-bottom: 2rem;
-        }
-
-        .input-group {
-            display: flex;
-            flex-direction: column;
-            gap: 0.5rem;
-        }
-
-        label {
-            font-size: 0.875rem;
-            font-weight: 500;
-        }
-
-        input,
-        select {
-            padding: 0.75rem;
-            border-radius: 0.5rem;
-            border: 1px solid rgba(255, 255, 255, 0.1);
-            background-color: var(--color-background);
-            color: var(--color-text);
-            font-size: 1rem;
-        }
-
-        button {
-            padding: 1rem 2rem;
-            border-radius: 0.5rem;
-            border: none;
-            background-color: var(--color-accent);
-            color: white;
-            font-weight: 600;
-            cursor: pointer;
-            transition: all 0.2s ease;
-        }
-
-        button:hover {
-            opacity: 0.9;
-            transform: translateY(-1px);
-        }
-
-        .icon-with-spinner {
-            display: flex;
-            align-items: center;
-            justify-content: center;
-            gap: 12px;
-            min-width: 180px;
-        }
-
-        .spinner {
-            width: 20px;
-            height: 20px;
-            border: 2px solid white;
-            border-top-color: transparent;
-            border-radius: 50%;
-            animation: spin 1s linear infinite;
-            flex-shrink: 0;
-        }
-
-        @keyframes spin {
-            to {
-                transform: rotate(360deg);
-            }
-        }
-
-        .pulse-container {
-            display: flex;
-            align-items: center;
-            justify-content: center;
-            gap: 12px;
-            min-width: 180px;
-        }
-
-        .pulse-circle {
-            width: 20px;
-            height: 20px;
-            border-radius: 50%;
-            background-color: white;
-            opacity: 0.2;
-            flex-shrink: 0;
-            transform: translateX(-0%) scale(var(--audio-level, 1));
-            transition: transform 0.1s ease;
-        }
-
-        /* Add styles for toast notifications */
-        .toast {
-            position: fixed;
-            top: 20px;
-            left: 50%;
-            transform: translateX(-50%);
-            padding: 16px 24px;
-            border-radius: 4px;
-            font-size: 14px;
-            z-index: 1000;
-            display: none;
-            box-shadow: 0 2px 5px rgba(0, 0, 0, 0.2);
-        }
-
-        .toast.error {
-            background-color: #f44336;
-            color: white;
-        }
-
-        .toast.warning {
-            background-color: #ffd700;
-            color: black;
-        }
-    </style>
-</head>
-
-
-<body>
-    <!-- Add toast element after body opening tag -->
-    <div id="error-toast" class="toast"></div>
-    <div style="text-align: center">
-        <h1>Gemini Voice Chat</h1>
-        <p>Speak with Gemini using real-time audio streaming</p>
-        <p>
-            Get a Gemini API key
-            <a href="https://ai.google.dev/gemini-api/docs/api-key">here</a>
-        </p>
-    </div>
-    <div class="container">
-        <div class="controls">
-            <div class="input-group">
-                <label for="api-key">API Key</label>
-                <input type="password" id="api-key" placeholder="Enter your API key">
-            </div>
-            <div class="input-group">
-                <label for="voice">Voice</label>
-                <select id="voice">
-                    <option value="Puck">Puck</option>
-                    <option value="Charon">Charon</option>
-                    <option value="Kore">Kore</option>
-                    <option value="Fenrir">Fenrir</option>
-                    <option value="Aoede">Aoede</option>
-                </select>
-            </div>
-        </div>
-
-        <div class="wave-container">
-            <div class="box-container">
-                <!-- Boxes will be dynamically added here -->
-            </div>
-        </div>
-
-        <button id="start-button">Start Recording</button>
-    </div>
-
-    <audio id="audio-output"></audio>
-
-    <script>
-        let peerConnection;
-        let audioContext;
-        let dataChannel;
-        let isRecording = false;
-        let webrtc_id;
-        let animationId;
-        let analyser;
-        let analyser_input;
-        let dataArray;
-        let dataArray_input;
-
-        const startButton = document.getElementById('start-button');
-        const apiKeyInput = document.getElementById('api-key');
-        const voiceSelect = document.getElementById('voice');
-        const audioOutput = document.getElementById('audio-output');
-        const boxContainer = document.querySelector('.box-container');
-
-        const numBars = 32;
-        for (let i = 0; i < numBars; i++) {
-            const box = document.createElement('div');
-            box.className = 'box';
-            boxContainer.appendChild(box);
-        }
-
-        function updateButtonState() {
-            if (peerConnection && (peerConnection.connectionState === 'connecting' || peerConnection.connectionState === 'new')) {
-                startButton.innerHTML = `
-                    <div class="icon-with-spinner">
-                        <div class="spinner"></div>
-                        <span>Connecting...</span>
-                    </div>
-                `;
-            } else if (peerConnection && peerConnection.connectionState === 'connected') {
-                startButton.innerHTML = `
-                    <div class="pulse-container">
-                        <div class="pulse-circle"></div>
-                        <span>Stop Recording</span>
-                    </div>
-                `;
-            } else {
-                startButton.innerHTML = 'Start Recording';
-            }
-        }
-
-        function showError(message) {
-            const toast = document.getElementById('error-toast');
-            toast.textContent = message;
-            toast.className = 'toast error';
-            toast.style.display = 'block';
-
-            // Hide toast after 5 seconds
-            setTimeout(() => {
-                toast.style.display = 'none';
-            }, 5000);
-        }
-
-        async function setupWebRTC() {
-            // Replace placeholder with actual configuration
-            const config = {
-                iceServers: [
-                    { urls: 'stun:stun.l.google.com:19302' }
-                ]
-            };
-            peerConnection = new RTCPeerConnection(config);
-            webrtc_id = Math.random().toString(36).substring(7);
-
-            const timeoutId = setTimeout(() => {
-                const toast = document.getElementById('error-toast');
-                toast.textContent = "Connection is taking longer than usual. Are you on a VPN?";
-                toast.className = 'toast warning';
-                toast.style.display = 'block';
-
-                // Hide warning after 5 seconds
-                setTimeout(() => {
-                    toast.style.display = 'none';
-                }, 5000);
-            }, 5000);
-
-            try {
-                // Check if mediaDevices is supported
-                if (!navigator.mediaDevices || !navigator.mediaDevices.getUserMedia) {
-                    const errorMsg = 'Your browser does not support audio recording. Please use Chrome, Firefox, or Edge.';
-                    showError(errorMsg);
-                    throw new Error(errorMsg);
-                }
-                
-                const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
-                stream.getTracks().forEach(track => peerConnection.addTrack(track, stream));
-
-                // Update audio visualization setup
-                audioContext = new AudioContext();
-                analyser_input = audioContext.createAnalyser();
-                const source = audioContext.createMediaStreamSource(stream);
-                source.connect(analyser_input);
-                analyser_input.fftSize = 64;
-                dataArray_input = new Uint8Array(analyser_input.frequencyBinCount);
-
-                function updateAudioLevel() {
-                    analyser_input.getByteFrequencyData(dataArray_input);
-                    const average = Array.from(dataArray_input).reduce((a, b) => a + b, 0) / dataArray_input.length;
-                    const audioLevel = average / 255;
-
-                    const pulseCircle = document.querySelector('.pulse-circle');
-                    if (pulseCircle) {
-                        console.log("audioLevel", audioLevel);
-                        pulseCircle.style.setProperty('--audio-level', 1 + audioLevel);
-                    }
-
-                    animationId = requestAnimationFrame(updateAudioLevel);
-                }
-                updateAudioLevel();
-
-                // Add connection state change listener
-                peerConnection.addEventListener('connectionstatechange', () => {
-                    console.log('connectionstatechange', peerConnection.connectionState);
-                    if (peerConnection.connectionState === 'connected') {
-                        clearTimeout(timeoutId);
-                        const toast = document.getElementById('error-toast');
-                        toast.style.display = 'none';
-                    }
-                    updateButtonState();
-                });
-
-                // Handle incoming audio
-                peerConnection.addEventListener('track', (evt) => {
-                    if (audioOutput && audioOutput.srcObject !== evt.streams[0]) {
-                        audioOutput.srcObject = evt.streams[0];
-                        audioOutput.play();
-
-                        // Set up audio visualization on the output stream
-                        audioContext = new AudioContext();
-                        analyser = audioContext.createAnalyser();
-                        const source = audioContext.createMediaStreamSource(evt.streams[0]);
-                        source.connect(analyser);
-                        analyser.fftSize = 2048;
-                        dataArray = new Uint8Array(analyser.frequencyBinCount);
-                        updateVisualization();
-                    }
-                });
-
-                // Create data channel for messages
-                dataChannel = peerConnection.createDataChannel('text');
-                dataChannel.onmessage = (event) => {
-                    const eventJson = JSON.parse(event.data);
-                    if (eventJson.type === "error") {
-                        showError(eventJson.message);
-                    } else if (eventJson.type === "send_input") {
-                        fetch('/input_hook', {
-                            method: 'POST',
-                            headers: {
-                                'Content-Type': 'application/json',
-                            },
-                            body: JSON.stringify({
-                                webrtc_id: webrtc_id,
-                                api_key: apiKeyInput.value,
-                                voice_name: voiceSelect.value
-                            })
-                        });
-                    }
-                };
-
-                // Create and send offer
-                const offer = await peerConnection.createOffer();
-                await peerConnection.setLocalDescription(offer);
-
-                await new Promise((resolve) => {
-                    if (peerConnection.iceGatheringState === "complete") {
-                        resolve();
-                    } else {
-                        const checkState = () => {
-                            if (peerConnection.iceGatheringState === "complete") {
-                                peerConnection.removeEventListener("icegatheringstatechange", checkState);
-                                resolve();
-                            }
-                        };
-                        peerConnection.addEventListener("icegatheringstatechange", checkState);
-                    }
-                });
-
-                const response = await fetch('/webrtc/offer', {
-                    method: 'POST',
-                    headers: { 'Content-Type': 'application/json' },
-                    body: JSON.stringify({
-                        sdp: peerConnection.localDescription.sdp,
-                        type: peerConnection.localDescription.type,
-                        webrtc_id: webrtc_id,
-                    })
-                });
-
-                const serverResponse = await response.json();
-
-                if (serverResponse.status === 'failed') {
-                    showError(serverResponse.meta.error === 'concurrency_limit_reached'
-                        ? `Too many connections. Maximum limit is ${serverResponse.meta.limit}`
-                        : serverResponse.meta.error);
-                    stopWebRTC();
-                    startButton.textContent = 'Start Recording';
-                    return;
-                }
-
-                await peerConnection.setRemoteDescription(serverResponse);
-            } catch (err) {
-                clearTimeout(timeoutId);
-                console.error('Error setting up WebRTC:', err);
-                showError('Failed to establish connection. Please try again.');
-                stopWebRTC();
-                startButton.textContent = 'Start Recording';
-            }
-        }
-
-        function updateVisualization() {
-            if (!analyser) return;
-
-            analyser.getByteFrequencyData(dataArray);
-            const bars = document.querySelectorAll('.box');
-
-            for (let i = 0; i < bars.length; i++) {
-                const barHeight = (dataArray[i] / 255) * 2;
-                bars[i].style.transform = `scaleY(${Math.max(0.1, barHeight)})`;
-            }
-
-            animationId = requestAnimationFrame(updateVisualization);
-        }
-
-        function stopWebRTC() {
-            if (peerConnection) {
-                // Clean up all transceivers and tracks
-                if (peerConnection.getTransceivers) {
-                    peerConnection.getTransceivers().forEach(transceiver => {
-                        if (transceiver.stop) {
-                            transceiver.stop();
-                        }
-                    });
-                }
-
-                if (peerConnection.getSenders) {
-                    peerConnection.getSenders().forEach(sender => {
-                        if (sender.track && sender.track.stop) sender.track.stop();
-                    });
-                }
-                
-                peerConnection.close();
-                peerConnection = null;
-            }
-            
-            if (animationId) {
-                cancelAnimationFrame(animationId);
-                animationId = null;
-            }
-            
-            if (audioContext) {
-                audioContext.close();
-                audioContext = null;
-            }
-            
-            updateButtonState();
-        }
-
-        startButton.addEventListener('click', () => {
-            // First check for browser support
-            if (!navigator.mediaDevices || !navigator.mediaDevices.getUserMedia) {
-                showError('Your browser does not support audio recording. Please use Chrome, Firefox, or Edge.');
-                return;
-            }
-            
-            if (!isRecording) {
-                setupWebRTC();
-                startButton.classList.add('recording');
-            } else {
-                stopWebRTC();
-                startButton.classList.remove('recording');
-            }
-            isRecording = !isRecording;
-        });
-    </script>
-</body>
-
-</html>