mirror of
https://github.com/HumanAIGC-Engineering/gradio-webrtc.git
synced 2026-02-05 18:09:23 +08:00
837330dcd8
* Turn integration * Add code: * type hint * Fix typehint * add code * format * WIP * trickle ice * bump version * Better docs * Modify * code * Mute icon for whisper * Add code * llama 4 demo * code * OpenAI interruptions * fix docs
616 lines
22 KiB
HTML
616 lines
22 KiB
HTML
<!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;
|
|
display: flex;
|
|
align-items: center;
|
|
justify-content: center;
|
|
gap: 12px;
|
|
min-width: 180px;
|
|
}
|
|
|
|
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;
|
|
}
|
|
|
|
.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;
|
|
}
|
|
|
|
/* Add styles for the mute toggle */
|
|
.mute-toggle {
|
|
width: 24px;
|
|
height: 24px;
|
|
cursor: pointer;
|
|
flex-shrink: 0;
|
|
}
|
|
|
|
.mute-toggle svg {
|
|
display: block;
|
|
}
|
|
|
|
#start-button {
|
|
margin-left: auto;
|
|
margin-right: auto;
|
|
}
|
|
</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 isMuted = false;
|
|
let analyser_input, dataArray_input;
|
|
let analyser, dataArray;
|
|
let source_input = null;
|
|
let source_output = null;
|
|
|
|
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);
|
|
}
|
|
|
|
// SVG Icons
|
|
const micIconSVG = `
|
|
<svg xmlns="http://www.w3.org/2000/svg" width="100%" height="100%" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
|
|
<path d="M12 1a3 3 0 0 0-3 3v8a3 3 0 0 0 6 0V4a3 3 0 0 0-3-3z"></path>
|
|
<path d="M19 10v2a7 7 0 0 1-14 0v-2"></path>
|
|
<line x1="12" y1="19" x2="12" y2="23"></line>
|
|
<line x1="8" y1="23" x2="16" y2="23"></line>
|
|
</svg>`;
|
|
|
|
const micMutedIconSVG = `
|
|
<svg xmlns="http://www.w3.org/2000/svg" width="100%" height="100%" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
|
|
<path d="M12 1a3 3 0 0 0-3 3v8a3 3 0 0 0 6 0V4a3 3 0 0 0-3-3z"></path>
|
|
<path d="M19 10v2a7 7 0 0 1-14 0v-2"></path>
|
|
<line x1="12" y1="19" x2="12" y2="23"></line>
|
|
<line x1="8" y1="23" x2="16" y2="23"></line>
|
|
<line x1="1" y1="1" x2="23" y2="23"></line>
|
|
</svg>`;
|
|
|
|
function updateButtonState() {
|
|
startButton.innerHTML = '';
|
|
startButton.onclick = null;
|
|
|
|
if (peerConnection && (peerConnection.connectionState === 'connecting' || peerConnection.connectionState === 'new')) {
|
|
startButton.innerHTML = `
|
|
<div class="icon-with-spinner">
|
|
<div class="spinner"></div>
|
|
<span>Connecting...</span>
|
|
</div>
|
|
`;
|
|
startButton.disabled = true;
|
|
} else if (peerConnection && peerConnection.connectionState === 'connected') {
|
|
const pulseContainer = document.createElement('div');
|
|
pulseContainer.className = 'pulse-container';
|
|
pulseContainer.innerHTML = `
|
|
<div class="pulse-circle"></div>
|
|
<span>Stop Recording</span>
|
|
`;
|
|
|
|
const muteToggle = document.createElement('div');
|
|
muteToggle.className = 'mute-toggle';
|
|
muteToggle.title = isMuted ? 'Unmute' : 'Mute';
|
|
muteToggle.innerHTML = isMuted ? micMutedIconSVG : micIconSVG;
|
|
muteToggle.addEventListener('click', toggleMute);
|
|
|
|
startButton.appendChild(pulseContainer);
|
|
startButton.appendChild(muteToggle);
|
|
startButton.disabled = false;
|
|
|
|
} else {
|
|
startButton.innerHTML = 'Start Recording';
|
|
startButton.disabled = false;
|
|
}
|
|
}
|
|
|
|
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);
|
|
}
|
|
|
|
function toggleMute(event) {
|
|
event.stopPropagation();
|
|
if (!peerConnection || peerConnection.connectionState !== 'connected') return;
|
|
|
|
isMuted = !isMuted;
|
|
console.log("Mute toggled:", isMuted);
|
|
|
|
peerConnection.getSenders().forEach(sender => {
|
|
if (sender.track && sender.track.kind === 'audio') {
|
|
sender.track.enabled = !isMuted;
|
|
console.log(`Audio track ${sender.track.id} enabled: ${!isMuted}`);
|
|
}
|
|
});
|
|
|
|
updateButtonState();
|
|
}
|
|
|
|
async function setupWebRTC() {
|
|
const config = __RTC_CONFIGURATION__;
|
|
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 {
|
|
const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
|
|
stream.getTracks().forEach(track => peerConnection.addTrack(track, stream));
|
|
|
|
if (!audioContext || audioContext.state === 'closed') {
|
|
audioContext = new AudioContext();
|
|
}
|
|
if (source_input) {
|
|
try { source_input.disconnect(); } catch (e) { console.warn("Error disconnecting previous input source:", e); }
|
|
source_input = null;
|
|
}
|
|
source_input = audioContext.createMediaStreamSource(stream);
|
|
analyser_input = audioContext.createAnalyser();
|
|
source_input.connect(analyser_input);
|
|
analyser_input.fftSize = 64;
|
|
dataArray_input = new Uint8Array(analyser_input.frequencyBinCount);
|
|
updateAudioLevel();
|
|
|
|
peerConnection.addEventListener('connectionstatechange', () => {
|
|
console.log('connectionstatechange', peerConnection.connectionState);
|
|
if (peerConnection.connectionState === 'connected') {
|
|
clearTimeout(timeoutId);
|
|
const toast = document.getElementById('error-toast');
|
|
toast.style.display = 'none';
|
|
if (analyser_input) updateAudioLevel();
|
|
if (analyser) updateVisualization();
|
|
} else if (['disconnected', 'failed', 'closed'].includes(peerConnection.connectionState)) {
|
|
// Explicitly stop animations if connection drops unexpectedly
|
|
// Note: stopWebRTC() handles the normal stop case
|
|
}
|
|
updateButtonState();
|
|
});
|
|
|
|
peerConnection.onicecandidate = ({ candidate }) => {
|
|
if (candidate) {
|
|
console.debug("Sending ICE candidate", candidate);
|
|
fetch('/webrtc/offer', {
|
|
method: 'POST',
|
|
headers: { 'Content-Type': 'application/json' },
|
|
body: JSON.stringify({
|
|
candidate: candidate.toJSON(),
|
|
webrtc_id: webrtc_id,
|
|
type: "ice-candidate",
|
|
})
|
|
})
|
|
}
|
|
};
|
|
|
|
peerConnection.addEventListener('track', (evt) => {
|
|
if (evt.track.kind === 'audio' && audioOutput) {
|
|
if (audioOutput.srcObject !== evt.streams[0]) {
|
|
audioOutput.srcObject = evt.streams[0];
|
|
audioOutput.play().catch(e => console.error("Audio play failed:", e));
|
|
|
|
if (!audioContext || audioContext.state === 'closed') {
|
|
console.warn("AudioContext not ready for output track analysis.");
|
|
return;
|
|
}
|
|
if (source_output) {
|
|
try { source_output.disconnect(); } catch (e) { console.warn("Error disconnecting previous output source:", e); }
|
|
source_output = null;
|
|
}
|
|
source_output = audioContext.createMediaStreamSource(evt.streams[0]);
|
|
analyser = audioContext.createAnalyser();
|
|
source_output.connect(analyser);
|
|
analyser.fftSize = 2048;
|
|
dataArray = new Uint8Array(analyser.frequencyBinCount);
|
|
updateVisualization();
|
|
}
|
|
}
|
|
});
|
|
|
|
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
|
|
})
|
|
});
|
|
}
|
|
};
|
|
|
|
const offer = await peerConnection.createOffer();
|
|
await peerConnection.setLocalDescription(offer);
|
|
|
|
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 ${serverResponse.meta.limit}`
|
|
: serverResponse.meta.error);
|
|
stopWebRTC();
|
|
startButton.textContent = 'Start Recording';
|
|
return;
|
|
}
|
|
|
|
await peerConnection.setRemoteDescription(serverResponse);
|
|
} catch (err) {
|
|
clearTimeout(timeoutId);
|
|
console.error('Error setting up WebRTC:', err);
|
|
showError('Failed to establish connection. Please try again.');
|
|
stopWebRTC();
|
|
startButton.textContent = 'Start Recording';
|
|
}
|
|
}
|
|
|
|
function updateVisualization() {
|
|
if (!analyser || !peerConnection || !['connected', 'connecting'].includes(peerConnection.connectionState)) {
|
|
const bars = document.querySelectorAll('.box');
|
|
bars.forEach(bar => bar.style.transform = 'scaleY(0.1)');
|
|
return;
|
|
}
|
|
|
|
analyser.getByteFrequencyData(dataArray);
|
|
const bars = document.querySelectorAll('.box');
|
|
|
|
for (let i = 0; i < bars.length; i++) {
|
|
const barHeight = (dataArray[i] / 255) * 2;
|
|
bars[i].style.transform = `scaleY(${Math.max(0.1, barHeight)})`;
|
|
}
|
|
|
|
requestAnimationFrame(updateVisualization);
|
|
}
|
|
|
|
function updateAudioLevel() {
|
|
if (!analyser_input || !peerConnection || !['connected', 'connecting'].includes(peerConnection.connectionState)) {
|
|
const pulseCircle = document.querySelector('.pulse-circle');
|
|
if (pulseCircle) {
|
|
pulseCircle.style.setProperty('--audio-level', 1);
|
|
}
|
|
return;
|
|
}
|
|
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) {
|
|
pulseCircle.style.setProperty('--audio-level', 1 + audioLevel);
|
|
}
|
|
|
|
requestAnimationFrame(updateAudioLevel);
|
|
}
|
|
|
|
function stopWebRTC() {
|
|
console.log("Running stopWebRTC");
|
|
if (peerConnection) {
|
|
peerConnection.getSenders().forEach(sender => {
|
|
if (sender.track) {
|
|
sender.track.stop();
|
|
}
|
|
});
|
|
peerConnection.ontrack = null;
|
|
peerConnection.onicegatheringstatechange = null;
|
|
peerConnection.onconnectionstatechange = null;
|
|
|
|
if (dataChannel) {
|
|
dataChannel.onmessage = null;
|
|
try { dataChannel.close(); } catch (e) { console.warn("Error closing data channel:", e); }
|
|
dataChannel = null;
|
|
}
|
|
try { peerConnection.close(); } catch (e) { console.warn("Error closing peer connection:", e); }
|
|
peerConnection = null;
|
|
}
|
|
|
|
if (audioOutput) {
|
|
audioOutput.pause();
|
|
audioOutput.srcObject = null;
|
|
}
|
|
|
|
if (source_input) {
|
|
try { source_input.disconnect(); } catch (e) { console.warn("Error disconnecting input source:", e); }
|
|
source_input = null;
|
|
}
|
|
if (source_output) {
|
|
try { source_output.disconnect(); } catch (e) { console.warn("Error disconnecting output source:", e); }
|
|
source_output = null;
|
|
}
|
|
|
|
if (audioContext && audioContext.state !== 'closed') {
|
|
audioContext.close().then(() => {
|
|
console.log("AudioContext closed successfully.");
|
|
audioContext = null;
|
|
}).catch(e => {
|
|
console.error("Error closing AudioContext:", e);
|
|
audioContext = null;
|
|
});
|
|
} else {
|
|
audioContext = null;
|
|
}
|
|
|
|
analyser_input = null;
|
|
dataArray_input = null;
|
|
analyser = null;
|
|
dataArray = null;
|
|
|
|
isMuted = false;
|
|
isRecording = false;
|
|
updateButtonState();
|
|
|
|
const bars = document.querySelectorAll('.box');
|
|
bars.forEach(bar => bar.style.transform = 'scaleY(0.1)');
|
|
const pulseCircle = document.querySelector('.pulse-circle');
|
|
if (pulseCircle) {
|
|
pulseCircle.style.setProperty('--audio-level', 1);
|
|
}
|
|
}
|
|
|
|
startButton.addEventListener('click', (event) => {
|
|
if (event.target.closest('.mute-toggle')) {
|
|
return;
|
|
}
|
|
|
|
if (peerConnection && peerConnection.connectionState === 'connected') {
|
|
console.log("Stop button clicked");
|
|
stopWebRTC();
|
|
} else if (!peerConnection || ['new', 'closed', 'failed', 'disconnected'].includes(peerConnection.connectionState)) {
|
|
console.log("Start button clicked");
|
|
if (!apiKeyInput.value) {
|
|
showError("Please enter your API Key.");
|
|
return;
|
|
}
|
|
setupWebRTC();
|
|
isRecording = true;
|
|
updateButtonState();
|
|
}
|
|
});
|
|
|
|
updateButtonState();
|
|
</script>
|
|
</body>
|
|
|
|
</html> |