Add example for "Talk to Azure OpenAi" (#181)

* Add example for "Talk to Azure OpenAi"

* Code

---------

Co-authored-by: Freddy Boulton <alfonsoboulton@gmail.com>

demo/talk_to_azure_openai/README.md

@@ -0,0 +1,15 @@
+---
+title: Talk to Azure OpenAI
+emoji: 🗣️
+colorFrom: purple
+colorTo: red
+sdk: gradio
+sdk_version: 5.16.0
+app_file: app.py
+pinned: false
+license: mit
+short_description: Talk to Azure OpenAI using their multimodal API
+tags: [webrtc, websocket, gradio, secret|TWILIO_ACCOUNT_SID, secret|TWILIO_AUTH_TOKEN, secret|OPENAI_API_KEY]
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

demo/talk_to_azure_openai/README_gradio.md

@@ -0,0 +1,15 @@
+---
+title: Talk to Azure OpenAI (Gradio UI)
+emoji: 🗣️
+colorFrom: purple
+colorTo: red
+sdk: gradio
+sdk_version: 5.16.0
+app_file: app.py
+pinned: false
+license: mit
+short_description: Talk to Azure OpenAI (Gradio UI)
+tags: [webrtc, websocket, gradio, secret|TWILIO_ACCOUNT_SID, secret|TWILIO_AUTH_TOKEN, secret|OPENAI_API_KEY]
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

demo/talk_to_azure_openai/app.py

@@ -0,0 +1,233 @@
+import asyncio
+import base64
+import json
+from pathlib import Path
+import sounddevice as sd
+
+import gradio as gr
+import numpy as np
+import aiohttp  # pip install aiohttp
+from dotenv import load_dotenv
+from fastapi import FastAPI
+from fastapi.responses import HTMLResponse, StreamingResponse
+from fastrtc import (
+    AdditionalOutputs,
+    AsyncStreamHandler,
+    Stream,
+    get_twilio_turn_credentials,
+    wait_for_item,
+)
+from gradio.utils import get_space
+
+load_dotenv()
+cur_dir = Path(__file__).parent
+load_dotenv("key.env")
+# sd.default.device = (3, 3)  # (Input-Gerät, Output-Gerät)
+
+# print(f"Used Mic: {sd.query_devices(3)['name']}")
+# print(f"Used Speaker: {sd.query_devices(3)['name']}")
+SAMPLE_RATE = 24000
+
+instruction = """
+<Role>
+You a helpful assistant.
+"""
+
+
+class AzureAudioHandler(AsyncStreamHandler):
+    def __init__(self) -> None:
+        super().__init__(
+            expected_layout="mono",
+            output_sample_rate=SAMPLE_RATE,
+            output_frame_size=480,
+            input_sample_rate=SAMPLE_RATE,
+        )
+        self.ws = None
+        self.session = None
+        self.output_queue = asyncio.Queue()
+        # This internal buffer is not used directly in receive_messages.
+        # Instead, multiple audio chunks are collected in the emit() method.
+        # If needed, a continuous buffer can also be implemented here.
+        # self.audio_buffer = bytearray()
+
+    def copy(self):
+        return AzureAudioHandler()
+
+    async def start_up(self):
+        """Connects to the Azure Real-time Audio API via WebSocket using aiohttp."""
+        # Replace the following placeholders with your actual Azure values:
+        azure_api_key = "your-api-key"  # e.g., "your-api-key"
+        azure_resource_name = "your-resource-name"  # e.g., "aigdopenai"
+        deployment_id = "your-deployment-id"  # e.g., "gpt-4o-realtime-preview"
+        api_version = "2024-10-01-preview"
+        azure_endpoint = (
+            f"wss://{azure_resource_name}.openai.azure.com/openai/realtime"
+            f"?api-version={api_version}&deployment={deployment_id}"
+        )
+        headers = {"api-key": azure_api_key}
+
+        self.session = aiohttp.ClientSession()
+        self.ws = await self.session.ws_connect(azure_endpoint, headers=headers)
+        # Send initial session parameters
+        session_update_message = {
+            "type": "session.update",
+            "session": {
+                "turn_detection": {"type": "server_vad"},
+                "instructions": instruction,
+                "voice": "ballad",  # Possible voices see  https://platform.openai.com/docs/guides/realtime-model-capabilities#voice-options
+            },
+        }
+
+        await self.ws.send_str(json.dumps(session_update_message))
+        # Start receiving messages asynchronously
+        asyncio.create_task(self.receive_messages())
+
+    async def receive_messages(self):
+        """Handles incoming WebSocket messages and processes them accordingly."""
+        async for msg in self.ws:
+            if msg.type == aiohttp.WSMsgType.TEXT:
+                print("Received event:", msg.data)  # Debug output
+                event = json.loads(msg.data)
+                event_type = event.get("type")
+                if event_type in ["final", "response.audio_transcript.done"]:
+                    transcript = event.get("transcript", "")
+
+                    # Wrap the transcript in an object with a .transcript attribute
+                    class TranscriptEvent:
+                        pass
+
+                    te = TranscriptEvent()
+                    te.transcript = transcript
+                    await self.output_queue.put(AdditionalOutputs(te))
+                elif event_type == "partial":
+                    print("Partial transcript:", event.get("transcript", ""))
+                elif event_type == "response.audio.delta":
+                    audio_message = event.get("delta")
+                    if audio_message:
+                        try:
+                            audio_bytes = base64.b64decode(audio_message)
+                            # Assuming 16-bit PCM (int16)
+                            audio_array = np.frombuffer(audio_bytes, dtype=np.int16)
+                            # Interpret as mono audio:
+                            audio_array = audio_array.reshape(1, -1)
+                            # Instead of playing the audio, add the chunk to the output queue
+                            await self.output_queue.put(
+                                (self.output_sample_rate, audio_array)
+                            )
+                        except Exception as e:
+                            print("Error processing audio data:", e)
+                else:
+                    print("Unknown event:", event)
+            elif msg.type == aiohttp.WSMsgType.ERROR:
+                break
+
+    async def receive(self, frame: tuple[int, np.ndarray]) -> None:
+        """Sends received audio frames to the WebSocket."""
+        if not self.ws or self.ws.closed:
+            return
+        try:
+            _, array = frame
+            array = array.squeeze()
+            audio_message = base64.b64encode(array.tobytes()).decode("utf-8")
+            message = {"type": "input_audio_buffer.append", "audio": audio_message}
+            await self.ws.send_str(json.dumps(message))
+        except aiohttp.ClientConnectionError as e:
+            print("Connection closed while sending:", e)
+            return
+
+    async def emit(self) -> tuple[int, np.ndarray] | AdditionalOutputs | None:
+        """
+        Collects multiple audio chunks from the queue before returning them as a single contiguous audio array.
+        This helps smooth playback.
+        """
+        item = await wait_for_item(self.output_queue)
+        # If it's a transcript event, return it immediately.
+        if not isinstance(item, tuple):
+            return item
+        # Otherwise, it is an audio chunk (sample_rate, audio_array)
+        sample_rate, first_chunk = item
+        audio_chunks = [first_chunk]
+        # Define a minimum length (e.g., 0.1 seconds)
+        min_samples = int(SAMPLE_RATE * 0.1)  # 0.1 sec
+        # Collect more audio chunks until we have enough samples
+        while audio_chunks and audio_chunks[0].shape[1] < min_samples:
+            try:
+                extra = self.output_queue.get_nowait()
+                if isinstance(extra, tuple):
+                    _, chunk = extra
+                    audio_chunks.append(chunk)
+                else:
+                    # If it's not an audio chunk, put it back
+                    await self.output_queue.put(extra)
+                    break
+            except asyncio.QueueEmpty:
+                break
+        # Concatenate collected chunks along the time axis (axis=1)
+        full_audio = np.concatenate(audio_chunks, axis=1)
+        return (sample_rate, full_audio)
+
+    async def shutdown(self) -> None:
+        """Closes the WebSocket and session properly."""
+        if self.ws:
+            await self.ws.close()
+            self.ws = None
+        if self.session:
+            await self.session.close()
+            self.session = None
+
+
+def update_chatbot(chatbot: list[dict], response) -> list[dict]:
+    """Appends the AI assistant's transcript response to the chatbot messages."""
+    chatbot.append({"role": "assistant", "content": response.transcript})
+    return chatbot
+
+
+chatbot = gr.Chatbot(type="messages")
+latest_message = gr.Textbox(type="text", visible=False)
+stream = Stream(
+    AzureAudioHandler(),
+    mode="send-receive",
+    modality="audio",
+    additional_inputs=[chatbot],
+    additional_outputs=[chatbot],
+    additional_outputs_handler=update_chatbot,
+    rtc_configuration=get_twilio_turn_credentials() if get_space() else None,
+    concurrency_limit=5 if get_space() else None,
+    time_limit=90 if get_space() else None,
+)
+
+app = FastAPI()
+stream.mount(app)
+
+
+@app.get("/")
+async def _():
+    rtc_config = get_twilio_turn_credentials() if get_space() else None
+    html_content = (cur_dir / "index.html").read_text()
+    html_content = html_content.replace("__RTC_CONFIGURATION__", json.dumps(rtc_config))
+    return HTMLResponse(content=html_content)
+
+
+@app.get("/outputs")
+def _(webrtc_id: str):
+    async def output_stream():
+        import json
+
+        async for output in stream.output_stream(webrtc_id):
+            s = json.dumps({"role": "assistant", "content": output.args[0].transcript})
+            yield f"event: output\ndata: {s}\n\n"
+
+    return StreamingResponse(output_stream(), media_type="text/event-stream")
+
+
+if __name__ == "__main__":
+    import os
+
+    if (mode := os.getenv("MODE")) == "UI":
+        stream.ui.launch(server_port=7860)
+    elif mode == "PHONE":
+        stream.fastphone(host="0.0.0.0", port=7860)
+    else:
+        import uvicorn
+
+        uvicorn.run(app, host="0.0.0.0", port=7860)

demo/talk_to_azure_openai/index.html

@@ -0,0 +1,356 @@
+
+<!DOCTYPE html>
+<html lang="en">
+
+<head>
+    <meta charset="UTF-8">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Azure OpenAI Real-Time Chat</title>
+    <style>
+        body {
+            font-family: "SF Pro Display", -apple-system, BlinkMacSystemFont, sans-serif;
+            background-color: #0a0a0a;
+            color: #ffffff;
+            margin: 0;
+            padding: 20px;
+            height: 100vh;
+            box-sizing: border-box;
+        }
+        .container {
+            max-width: 800px;
+            margin: 0 auto;
+            height: calc(100% - 100px);
+        }
+        .logo {
+            text-align: center;
+            margin-bottom: 40px;
+        }
+        .chat-container {
+            border: 1px solid #333;
+            padding: 20px;
+            height: 90%;
+            box-sizing: border-box;
+            display: flex;
+            flex-direction: column;
+        }
+        .chat-messages {
+            flex-grow: 1;
+            overflow-y: auto;
+            margin-bottom: 20px;
+            padding: 10px;
+        }
+        .message {
+            margin-bottom: 20px;
+            padding: 12px;
+            border-radius: 4px;
+            font-size: 16px;
+            line-height: 1.5;
+        }
+        .message.user {
+            background-color: #1a1a1a;
+            margin-left: 20%;
+        }
+        .message.assistant {
+            background-color: #262626;
+            margin-right: 20%;
+        }
+        .controls {
+            text-align: center;
+            margin-top: 20px;
+        }
+        button {
+            background-color: transparent;
+            color: #ffffff;
+            border: 1px solid #ffffff;
+            padding: 12px 24px;
+            font-family: inherit;
+            font-size: 16px;
+            cursor: pointer;
+            transition: all 0.3s;
+            text-transform: uppercase;
+            letter-spacing: 1px;
+        }
+        button:hover {
+            border-width: 2px;
+            transform: scale(1.02);
+            box-shadow: 0 0 10px rgba(255, 255, 255, 0.2);
+        }
+        #audio-output {
+            display: none;
+        }
+        .icon-with-spinner {
+            display: flex;
+            align-items: center;
+            justify-content: center;
+            gap: 12px;
+            min-width: 180px;
+        }
+        .spinner {
+            width: 20px;
+            height: 20px;
+            border: 2px solid #ffffff;
+            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: #ffffff;
+            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 class="container">
+        <div class="logo">
+            <h1>OpenAI Real-Time Chat</h1>
+        </div>
+        <div class="chat-container">
+            <div class="chat-messages" id="chat-messages"></div>
+        </div>
+        <div class="controls">
+            <button id="start-button">Start Conversation</button>
+        </div>
+    </div>
+    <audio id="audio-output"></audio>
+
+    <script>
+        let peerConnection;
+        let webrtc_id;
+        const audioOutput = document.getElementById('audio-output');
+        const startButton = document.getElementById('start-button');
+        const chatMessages = document.getElementById('chat-messages');
+        let audioLevel = 0;
+        let animationFrame;
+        let audioContext, analyser, audioSource;
+        function updateButtonState() {
+            const button = document.getElementById('start-button');
+            if (peerConnection && (peerConnection.connectionState === 'connecting' || peerConnection.connectionState === 'new')) {
+                button.innerHTML = `
+                    <div class="icon-with-spinner">
+                        <div class="spinner"></div>
+                        <span>Connecting...</span>
+                    </div>
+                `;
+            } else if (peerConnection && peerConnection.connectionState === 'connected') {
+                button.innerHTML = `
+                    <div class="pulse-container">
+                        <div class="pulse-circle"></div>
+                        <span>Stop Conversation</span>
+                    </div>
+                `;
+            } else {
+                button.innerHTML = 'Start Conversation';
+            }
+        }
+        function setupAudioVisualization(stream) {
+            audioContext = new (window.AudioContext || window.webkitAudioContext)();
+            analyser = audioContext.createAnalyser();
+            audioSource = audioContext.createMediaStreamSource(stream);
+            audioSource.connect(analyser);
+            analyser.fftSize = 64;
+            const dataArray = new Uint8Array(analyser.frequencyBinCount);
+            function updateAudioLevel() {
+                analyser.getByteFrequencyData(dataArray);
+                const average = Array.from(dataArray).reduce((a, b) => a + b, 0) / dataArray.length;
+                audioLevel = average / 255;
+                // Update CSS variable instead of rebuilding the button
+                const pulseCircle = document.querySelector('.pulse-circle');
+                if (pulseCircle) {
+                    pulseCircle.style.setProperty('--audio-level', 1 + audioLevel);
+                }
+                animationFrame = requestAnimationFrame(updateAudioLevel);
+            }
+            updateAudioLevel();
+        }
+        function showError(message) {
+            const toast = document.getElementById('error-toast');
+            toast.textContent = message;
+            toast.style.display = 'block';
+            // Hide toast after 5 seconds
+            setTimeout(() => {
+                toast.style.display = 'none';
+            }, 5000);
+        }
+        async function setupWebRTC() {
+            isConnecting = true;
+            const config = __RTC_CONFIGURATION__;
+            peerConnection = new RTCPeerConnection(config);
+            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 {
+                const stream = await navigator.mediaDevices.getUserMedia({
+                    audio: true
+                });
+                setupAudioVisualization(stream);
+                stream.getTracks().forEach(track => {
+                    peerConnection.addTrack(track, stream);
+                });
+                peerConnection.addEventListener('track', (evt) => {
+                    if (audioOutput.srcObject !== evt.streams[0]) {
+                        audioOutput.srcObject = evt.streams[0];
+                        audioOutput.play();
+                    }
+                });
+                const dataChannel = peerConnection.createDataChannel('text');
+                dataChannel.onmessage = (event) => {
+                    const eventJson = JSON.parse(event.data);
+                    if (eventJson.type === "error") {
+                        showError(eventJson.message);
+                    }
+                };
+                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);
+                    }
+                });
+                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();
+                });
+                webrtc_id = Math.random().toString(36).substring(7);
+                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);
+                    stop();
+                    return;
+                }
+                await peerConnection.setRemoteDescription(serverResponse);
+                const eventSource = new EventSource('/outputs?webrtc_id=' + webrtc_id);
+                eventSource.addEventListener("output", (event) => {
+                    const eventJson = JSON.parse(event.data);
+                    addMessage("assistant", eventJson.content);
+                });
+            } catch (err) {
+                clearTimeout(timeoutId);
+                console.error('Error setting up WebRTC:', err);
+                showError('Failed to establish connection. Please try again.');
+                stop();
+            }
+        }
+        function addMessage(role, content) {
+            const messageDiv = document.createElement('div');
+            messageDiv.classList.add('message', role);
+            messageDiv.textContent = content;
+            chatMessages.appendChild(messageDiv);
+            chatMessages.scrollTop = chatMessages.scrollHeight;
+        }
+        function stop() {
+            if (animationFrame) {
+                cancelAnimationFrame(animationFrame);
+            }
+            if (audioContext) {
+                audioContext.close();
+                audioContext = null;
+                analyser = null;
+                audioSource = null;
+            }
+            if (peerConnection) {
+                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();
+                    });
+                }
+                console.log('closing');
+                peerConnection.close();
+            }
+            updateButtonState();
+            audioLevel = 0;
+        }
+        startButton.addEventListener('click', () => {
+            console.log('clicked');
+            console.log(peerConnection, peerConnection?.connectionState);
+            if (!peerConnection || peerConnection.connectionState !== 'connected') {
+                setupWebRTC();
+            } else {
+                console.log('stopping');
+                stop();
+            }
+        });
+    </script>
+</body>
+
+</html>

demo/talk_to_azure_openai/requirements.txt

@@ -0,0 +1,123 @@
+aiofiles==23.2.1
+aiohappyeyeballs==2.6.1
+aiohttp==3.11.13
+aiohttp-retry==2.9.1
+aioice==0.9.0
+aiortc==1.10.1
+aiosignal==1.3.2
+annotated-types==0.7.0
+anyio==4.8.0
+attrs==25.2.0
+audioread==3.0.1
+av==13.1.0
+babel==2.17.0
+certifi==2025.1.31
+cffi==1.17.1
+charset-normalizer==3.4.1
+click==8.1.8
+colorama==0.4.6
+coloredlogs==15.0.1
+colorlog==6.9.0
+cryptography==44.0.2
+csvw==3.5.1
+decorator==5.2.1
+distro==1.9.0
+dlinfo==2.0.0
+dnspython==2.7.0
+espeakng-loader==0.2.4
+fastapi==0.115.11
+fastrtc==0.0.14
+ffmpy==0.5.0
+filelock==3.17.0
+flatbuffers==25.2.10
+frozenlist==1.5.0
+fsspec==2025.3.0
+google-crc32c==1.6.0
+gradio==5.20.1
+gradio_client==1.7.2
+groovy==0.1.2
+h11==0.14.0
+httpcore==1.0.7
+httpx==0.28.1
+huggingface-hub==0.29.3
+humanfriendly==10.0
+idna==3.10
+ifaddr==0.2.0
+isodate==0.7.2
+Jinja2==3.1.6
+jiter==0.9.0
+joblib==1.4.2
+jsonschema==4.23.0
+jsonschema-specifications==2024.10.1
+kokoro-onnx==0.4.5
+language-tags==1.2.0
+lazy_loader==0.4
+librosa==0.11.0
+llvmlite==0.44.0
+markdown-it-py==3.0.0
+MarkupSafe==2.1.5
+mdurl==0.1.2
+mpmath==1.3.0
+msgpack==1.1.0
+multidict==6.1.0
+numba==0.61.0
+numpy==2.1.3
+onnxruntime==1.21.0
+openai==1.66.2
+orjson==3.10.15
+packaging==24.2
+pandas==2.2.3
+phonemizer-fork==3.3.1
+pillow==11.1.0
+platformdirs==4.3.6
+pooch==1.8.2
+propcache==0.3.0
+protobuf==6.30.0
+pycparser==2.22
+pydantic==2.10.6
+pydantic_core==2.27.2
+pydub==0.25.1
+pyee==12.1.1
+Pygments==2.19.1
+PyJWT==2.10.1
+pylibsrtp==0.11.0
+pyOpenSSL==25.0.0
+pyparsing==3.2.1
+python-dateutil==2.9.0.post0
+python-dotenv==1.0.1
+python-multipart==0.0.20
+pytz==2025.1
+PyYAML==6.0.2
+rdflib==7.1.3
+referencing==0.36.2
+regex==2024.11.6
+requests==2.32.3
+rfc3986==1.5.0
+rich==13.9.4
+rpds-py==0.23.1
+ruff==0.9.10
+safehttpx==0.1.6
+scikit-learn==1.6.1
+scipy==1.15.2
+segments==2.3.0
+semantic-version==2.10.0
+shellingham==1.5.4
+six==1.17.0
+sniffio==1.3.1
+sounddevice==0.5.1
+soundfile==0.13.1
+soxr==0.5.0.post1
+starlette==0.46.1
+sympy==1.13.3
+threadpoolctl==3.5.0
+tomlkit==0.13.2
+tqdm==4.67.1
+twilio==9.5.0
+typer==0.15.2
+typing_extensions==4.12.2
+tzdata==2025.1
+uritemplate==4.1.1
+urllib3==2.3.0
+uvicorn==0.34.0
+websockets==15.0.1
+yarl==1.18.3

docs/cookbook.md

@@ -35,6 +35,7 @@ A collection of applications built with FastRTC. Click on the tags below to find
   <button class="tag-button" data-tag="kyutai"><code>Kyutai</code></button>
   <button class="tag-button" data-tag="agentic"><code>Agentic</code></button>
   <button class="tag-button" data-tag="local"><code>Local Models</code></button>
+  <button class="tag-button" data-tag="electron"><code>Electron</code></button>
 </div>
 
 <script>
@@ -331,7 +332,7 @@ document.querySelectorAll('.tag-button').forEach(button => {
 
     [:octicons-code-16: Code](https://github.com/sofi444/realtime-transcription-fastrtc/blob/main/main.py)
 
-    -   :speaking_head:{ .lg .middle } __Talk to Claude - Electron App__
+-   :speaking_head:{ .lg .middle } __Talk to Claude - Electron App__
 {: data-tags="audio,electron"}
 
     ---
@@ -341,8 +342,16 @@ document.querySelectorAll('.tag-button').forEach(button => {
     <video width=98% src="https://github.com/user-attachments/assets/df4628e4-ef0f-4a78-ab9b-1ed2374b1cae" controls style="text-align: center"></video>
 
     [:octicons-arrow-right-24: Demo](https://github.com/swairshah/voice-agent)
-    
+
     [:octicons-code-16: Code](https://github.com/swairshah/voice-agent)
 
+-   :speaking_head:{ .lg .middle } __Azure Realtime API__
+{: data-tags="audio,real-time-api"}
+
+    ---
+
+    Use the Azure Realtime API to create a real-time voice chat with GPT-4o.
+
+    [:octicons-code-16: Code](https://github.com/freddyaboulton/fastrtc/tree/main/demo/talk_to_azure_openai)
 
 </div>