Add Web Speech Recognition option for real-time streaming

Provides two transcription options:

1. Browser Whisper (Offline, Batch) - RECOMMENDED
   - 100% offline, zero network calls
   - HIPAA-compliant, audio never leaves device
   - Highest accuracy (Whisper)
   - Processes after recording (batch mode)
   - Models self-hosted, bundled in v2

2. Web Speech API (Real-time, Streaming) - EXPERIMENTAL
   - Real-time transcription (see words as you speak)
   - Uses browser's built-in speech recognition
   - ⚠️ Sends audio to cloud (Chrome/Edge → Google)
   - ⚠️ NOT HIPAA-compliant
   - Requires user consent with clear warnings

Features:
- Settings UI for both options
- Clear privacy warnings for Web Speech
- Mutual exclusion (only one active at a time)
- Browser detection shows which provider is used
- Confirmation dialog before enabling Web Speech

Use Cases:
- Clinical/HIPAA: Use Browser Whisper only
- Personal/Non-clinical: Can use Web Speech for real-time feedback
- Maximum privacy: Browser Whisper (offline)
- Maximum speed: Web Speech (if privacy not required)

Implementation:
- speechRecognition.js: Web Speech API wrapper
- transcriptionSettings.js: Settings UI handler
- Privacy info displayed per browser

User can choose based on their privacy vs. speed preference.
This commit is contained in:
ifedan-ed 2026-03-31 21:57:24 +00:00
parent a7dd08c9d1
commit 196f4432f0
4 changed files with 352 additions and 0 deletions

View file

@ -61,6 +61,30 @@
<p style="font-size:11px;color:var(--orange);margin:4px 0 0;display:none;" id="browser-whisper-csp-warning"><i class="fas fa-exclamation-triangle"></i> <strong>Network/Firewall Issue:</strong> If model download fails, check that <code>cdn.jsdelivr.net</code> and <code>huggingface.co</code> are not blocked. Server transcription will be used as fallback.</p>
</div>
<!-- Web Speech Recognition (Real-time Streaming) -->
<div class="settings-section card" id="web-speech-section" style="border-left:3px solid var(--orange);">
<h3><i class="fas fa-wave-square"></i> Real-Time Streaming Transcription</h3>
<div style="background:var(--orange-light);padding:12px;border-radius:6px;margin-bottom:12px;">
<p style="font-size:13px;color:var(--orange-dark);margin:0;"><i class="fas fa-exclamation-triangle"></i> <strong>Privacy Warning:</strong> Uses your browser's built-in speech recognition, which <strong>may send audio to cloud servers</strong> (Chrome/Edge send to Google). Only enable if you accept this trade-off for real-time transcription.</p>
</div>
<p style="font-size:13px;color:var(--g600);">See words appear as you speak (streaming). Overrides browser and server transcription when enabled. <strong>Not HIPAA-compliant</strong> in most browsers.</p>
<div style="display:flex;align-items:center;gap:12px;flex-wrap:wrap;margin-bottom:12px;">
<label style="font-size:13px;font-weight:600;">Enable real-time streaming:</label>
<label style="display:flex;align-items:center;gap:6px;cursor:pointer;">
<input type="checkbox" id="web-speech-enabled" style="accent-color:var(--orange);width:16px;height:16px;">
<span style="font-size:13px;" id="web-speech-status">Off</span>
</label>
</div>
<div id="web-speech-privacy-info" style="font-size:12px;color:var(--g500);padding:10px;background:var(--g50);border-radius:6px;margin-top:8px;">
<p style="margin:0 0 4px;font-weight:600;">Current Browser:</p>
<p style="margin:0;" id="web-speech-browser-info">Detecting...</p>
</div>
<p style="font-size:11px;color:var(--g400);margin:8px 0 0;"><i class="fas fa-info-circle"></i> <strong>Trade-off:</strong> Immediate transcription vs. privacy. For maximum privacy, use Browser Whisper (offline batch mode) or Server transcription with HIPAA-eligible provider.</p>
</div>
<!-- 2FA -->
<div class="settings-section card">
<h3><i class="fas fa-shield-halved"></i> Two-Factor Authentication</h3>

View file

@ -303,6 +303,8 @@
<script defer src="/js/audioBackup.js"></script>
<script defer src="/js/correctionTracker.js"></script>
<script defer src="/js/browserWhisper.js"></script>
<script defer src="/js/speechRecognition.js"></script>
<script defer src="/js/transcriptionSettings.js"></script>
<script defer src="/js/voicePreferences.js"></script>
<script defer src="/js/app.js"></script>
<script defer src="/js/auth.js"></script>

View file

@ -0,0 +1,160 @@
// ============================================================
// WEB SPEECH RECOGNITION — Real-time streaming transcription
// ⚠️ PRIVACY WARNING: Uses browser's built-in ASR which may
// send audio to cloud servers (Chrome/Edge → Google servers)
// Only use if you accept this trade-off for real-time streaming
// ============================================================
(function() {
var STORAGE_ENABLED = 'ped_web_speech_enabled';
var SpeechRecognition = window.SpeechRecognition || window.webkitSpeechRecognition;
var _recognition = null;
var _transcript = '';
var _isListening = false;
var _onPartialCallback = null;
var _onFinalCallback = null;
window.WebSpeechRecognition = {
isSupported: function() {
return typeof SpeechRecognition !== 'undefined';
},
isEnabled: function() {
try {
return localStorage.getItem(STORAGE_ENABLED) === '1';
} catch(e) {
return false;
}
},
setEnabled: function(val) {
try {
localStorage.setItem(STORAGE_ENABLED, val ? '1' : '0');
} catch(e) {}
},
startListening: function(options) {
if (!this.isSupported()) {
return Promise.reject(new Error('Web Speech API not supported'));
}
if (!this.isEnabled()) {
return Promise.reject(new Error('Web Speech Recognition not enabled'));
}
_transcript = '';
_isListening = true;
var opts = options || {};
_onPartialCallback = opts.onPartial || function() {};
_onFinalCallback = opts.onFinal || function() {};
_recognition = new SpeechRecognition();
_recognition.continuous = true; // Keep listening
_recognition.interimResults = true; // Show partial results
_recognition.lang = opts.language || 'en-US';
_recognition.maxAlternatives = 1;
return new Promise(function(resolve, reject) {
_recognition.onstart = function() {
console.log('[WebSpeech] Started listening');
};
_recognition.onresult = function(event) {
var interim = '';
var final = '';
for (var i = event.resultIndex; i < event.results.length; i++) {
var transcript = event.results[i][0].transcript;
if (event.results[i].isFinal) {
final += transcript + ' ';
_transcript += transcript + ' ';
} else {
interim += transcript;
}
}
// Callback with partial results (shown in real-time)
if (interim && _onPartialCallback) {
_onPartialCallback(interim);
}
// Callback with final results (confirmed words)
if (final && _onFinalCallback) {
_onFinalCallback(final.trim());
}
};
_recognition.onerror = function(event) {
console.error('[WebSpeech] Error:', event.error);
_isListening = false;
if (event.error === 'no-speech') {
reject(new Error('No speech detected'));
} else if (event.error === 'not-allowed') {
reject(new Error('Microphone permission denied'));
} else {
reject(new Error('Speech recognition error: ' + event.error));
}
};
_recognition.onend = function() {
_isListening = false;
console.log('[WebSpeech] Stopped listening');
resolve(_transcript.trim());
};
try {
_recognition.start();
} catch (e) {
reject(new Error('Failed to start recognition: ' + e.message));
}
});
},
stopListening: function() {
if (_recognition && _isListening) {
_recognition.stop();
}
},
isListening: function() {
return _isListening;
},
getCurrentTranscript: function() {
return _transcript.trim();
},
// Check if browser likely sends to cloud
getPrivacyInfo: function() {
var isChrome = /Chrome/.test(navigator.userAgent) && /Google Inc/.test(navigator.vendor);
var isEdge = /Edg/.test(navigator.userAgent);
if (isChrome || isEdge) {
return {
provider: 'Google Cloud Speech',
privacy: 'Audio sent to Google servers',
warning: 'NOT HIPAA-compliant'
};
} else if (/Safari/.test(navigator.userAgent) && !/Chrome/.test(navigator.userAgent)) {
return {
provider: 'Apple Speech Recognition',
privacy: 'May process on-device or Apple servers',
warning: 'Check Apple privacy policy'
};
} else {
return {
provider: 'Unknown',
privacy: 'May send audio to cloud servers',
warning: 'Privacy unknown'
};
}
}
};
})();

View file

@ -0,0 +1,166 @@
// ============================================================
// TRANSCRIPTION SETTINGS — Browser Whisper + Web Speech API
// ============================================================
(function() {
var _inited = false;
document.addEventListener('tabChanged', function(e) {
if (e.detail.tab !== 'settings' || _inited) return;
_inited = true;
initTranscriptionSettings();
});
// Also init on direct page load
if (document.readyState === 'loading') {
document.addEventListener('DOMContentLoaded', function() {
setTimeout(function() {
if (!_inited && document.getElementById('browser-whisper-enabled')) {
_inited = true;
initTranscriptionSettings();
}
}, 500);
});
} else {
setTimeout(function() {
if (!_inited && document.getElementById('browser-whisper-enabled')) {
_inited = true;
initTranscriptionSettings();
}
}, 500);
}
function initTranscriptionSettings() {
console.log('[TranscriptionSettings] Initializing...');
// ── Browser Whisper ──────────────────────────────────────
var whisperCheckbox = document.getElementById('browser-whisper-enabled');
var whisperStatus = document.getElementById('browser-whisper-status');
var whisperModel = document.getElementById('browser-whisper-model');
var whisperPreload = document.getElementById('btn-whisper-preload');
var whisperProgress = document.getElementById('browser-whisper-progress');
var whisperProgressText = document.getElementById('browser-whisper-progress-text');
if (whisperCheckbox && window.BrowserWhisper) {
whisperCheckbox.checked = BrowserWhisper.isEnabled();
whisperStatus.textContent = BrowserWhisper.isEnabled() ? 'On — audio stays on device' : 'Off';
if (whisperModel) {
whisperModel.value = BrowserWhisper.getModel();
}
whisperCheckbox.addEventListener('change', function() {
BrowserWhisper.setEnabled(whisperCheckbox.checked);
whisperStatus.textContent = whisperCheckbox.checked ? 'On — audio stays on device' : 'Off';
if (whisperCheckbox.checked) {
// Auto-preload on enable
BrowserWhisper.preload(function(file, pct) {
if (whisperProgress && whisperProgressText) {
if (pct >= 100) {
whisperProgress.style.display = 'none';
return;
}
whisperProgress.style.display = 'block';
whisperProgressText.textContent = file + (pct > 0 ? ' ' + pct + '%' : '');
}
});
}
});
if (whisperModel) {
whisperModel.addEventListener('change', function() {
BrowserWhisper.setModel(whisperModel.value);
});
}
if (whisperPreload) {
whisperPreload.addEventListener('click', function() {
if (!BrowserWhisper || !BrowserWhisper.isSupported()) {
showToast('Browser Whisper not supported', 'error');
return;
}
whisperPreload.disabled = true;
whisperPreload.innerHTML = '<i class="fas fa-spinner fa-spin"></i> Downloading...';
whisperProgress.style.display = 'block';
whisperProgressText.textContent = 'Initializing...';
BrowserWhisper.setEnabled(true);
whisperCheckbox.checked = true;
whisperStatus.textContent = 'On — audio stays on device';
BrowserWhisper.preload(function(file, pct) {
if (pct >= 100) {
whisperProgress.style.display = 'none';
whisperPreload.disabled = false;
whisperPreload.innerHTML = '<i class="fas fa-download"></i> Pre-download model';
showToast('Whisper model ready!', 'success');
return;
}
whisperProgress.style.display = 'block';
whisperProgressText.textContent = file + (pct > 0 ? ' ' + pct + '%' : '');
});
});
}
}
// ── Web Speech Recognition ───────────────────────────────
var speechCheckbox = document.getElementById('web-speech-enabled');
var speechStatus = document.getElementById('web-speech-status');
var speechBrowserInfo = document.getElementById('web-speech-browser-info');
if (speechCheckbox && window.WebSpeechRecognition) {
speechCheckbox.checked = WebSpeechRecognition.isEnabled();
speechStatus.textContent = WebSpeechRecognition.isEnabled() ? 'On — real-time streaming' : 'Off';
// Show privacy info
if (speechBrowserInfo) {
var privacyInfo = WebSpeechRecognition.getPrivacyInfo();
speechBrowserInfo.innerHTML =
'<strong>Provider:</strong> ' + privacyInfo.provider + '<br>' +
'<strong>Privacy:</strong> ' + privacyInfo.privacy + '<br>' +
'<strong style="color:var(--orange);">⚠️ ' + privacyInfo.warning + '</strong>';
}
speechCheckbox.addEventListener('change', function() {
if (speechCheckbox.checked) {
// Show confirmation for privacy warning
if (!confirm(
'WARNING: Real-time streaming uses your browser\'s speech recognition, ' +
'which may send audio to cloud servers (e.g., Google).\n\n' +
'This is NOT HIPAA-compliant.\n\n' +
'Only enable if you understand and accept this privacy trade-off.\n\n' +
'Continue?'
)) {
speechCheckbox.checked = false;
return;
}
// Disable Browser Whisper if enabling Web Speech
if (whisperCheckbox && whisperCheckbox.checked) {
whisperCheckbox.checked = false;
BrowserWhisper.setEnabled(false);
whisperStatus.textContent = 'Off';
showToast('Browser Whisper disabled (Web Speech takes priority)', 'info');
}
}
WebSpeechRecognition.setEnabled(speechCheckbox.checked);
speechStatus.textContent = speechCheckbox.checked ? 'On — real-time streaming' : 'Off';
});
// If not supported, disable and show message
if (!WebSpeechRecognition.isSupported()) {
speechCheckbox.disabled = true;
speechStatus.textContent = 'Not supported in this browser';
if (speechBrowserInfo) {
speechBrowserInfo.innerHTML = '<strong style="color:var(--red);">Not supported</strong> - Web Speech API not available in this browser.';
}
}
}
console.log('✅ Transcription settings initialized');
}
})();