auto: 2026-07-09_06:00 changes

1 files changed, 62 insertions, 29 deletions

diff --git a/modules/ext/voice.js b/modules/ext/voice.js
index b243ace..284d6f3 100644
--- a/modules/ext/voice.js
+++ b/modules/ext/voice.js
@@ -9,9 +9,29 @@ class VoiceModule {
         this._lastError = '';
         this._sttEndpoint = '/vmai-api/stt';  // VaultMind STT via nginx proxy
         this._recordingDuration = 10000;        // 10s auto-stop
+        this._voices = [];
+        this._autoSpeak = false;
+    }
+    init(terminal) {
+        this.terminal = terminal;
+        // Pre-load TTS voices (Chrome loads async)
+        if (typeof speechSynthesis !== 'undefined') {
+            this._voices = speechSynthesis.getVoices() || [];
+            if (this._voices.length === 0) {
+                var self = this;
+                speechSynthesis.onvoiceschanged = function() {
+                    self._voices = speechSynthesis.getVoices() || [];
+                };
+            }
+        }
+    }
+    _getBestVoice() {
+        var voices = this._voices.length > 0 ? this._voices : (typeof speechSynthesis !== 'undefined' ? speechSynthesis.getVoices() : []);
+        return voices.find(function(v) { return v.lang && v.lang.startsWith('en') && v.name.includes('Google'); })
+            || voices.find(function(v) { return v.lang && v.lang.startsWith('en'); })
+            || (voices.length > 0 ? voices[0] : null);
     }
     requirements() {
-        // MediaRecorder + getUserMedia work on ALL modern browsers
         const hasMic = typeof navigator !== 'undefined' && !!navigator.mediaDevices?.getUserMedia;
         const hasRecorder = typeof MediaRecorder !== 'undefined';
         if (hasMic && hasRecorder) {
@@ -23,26 +43,20 @@ class VoiceModule {
         return { met: false, missing };
     }
     getHelpInfo() {
-        return { 'voice': 'voice [stop] — Start voice recording (Whisper STT), or type "voice stop" to cancel' };
+        return { 'voice': 'voice [stop] — Start voice recording (Whisper STT), or type "voice stop" to cancel', 'speak': 'speak [text] — Text-to-speech. No args = toggle auto-speak for AI responses' };
     }
     registerCommands() {
         this.commands['voice'] = (args) => this.handle(args);
         this.commands['speak'] = (args) => this.speak(args);
     }
 
-    getHelpInfo() {
-        return { 'voice': 'voice [stop] — Start voice recording (Whisper STT), or type "voice stop" to cancel', 'speak': 'speak [text] — Text-to-speech. No args = toggle auto-speak for AI responses' };
-    }
-
     speak(args) {
         const text = args.join(' ').trim();
         if (!text) {
-            // If no text, check if speechSynthesis is available
             if (typeof speechSynthesis === 'undefined') {
                 this.terminal.print('🔊 TTS not supported in this browser', 'error');
                 return;
             }
-            // Toggle auto-speak mode
             this._autoSpeak = !this._autoSpeak;
             this.terminal.print(this._autoSpeak ? '🔊 Auto-speak ON — AI responses will be spoken' : '🔊 Auto-speak OFF', 'info');
             return;
@@ -51,40 +65,59 @@ class VoiceModule {
             this.terminal.print('🔊 TTS not supported in this browser', 'error');
             return;
         }
+        this._doSpeak(text);
+        this.terminal.print('🔊 Speaking: "' + text.substring(0, 80) + (text.length > 80 ? '...' : '') + '"', 'info');
+    }
+
+    _doSpeak(text) {
+        var self = this;
         speechSynthesis.cancel();
-        const utter = new SpeechSynthesisUtterance(text);
+        var utter = new SpeechSynthesisUtterance(text);
         utter.rate = parseFloat(localStorage.getItem('webx11_tts_rate') || '1.0');
         utter.pitch = 1.0;
         utter.volume = 1.0;
-        // Try to use a good English voice
-        const voices = speechSynthesis.getVoices();
-        const enVoice = voices.find(v => v.lang.startsWith('en') && v.name.includes('Google')) || voices.find(v => v.lang.startsWith('en'));
-        if (enVoice) utter.voice = enVoice;
-        utter.onstart = () => this._updateButton(true);
-        utter.onend = () => this._updateButton(false);
-        utter.onerror = () => this._updateButton(false);
-        speechSynthesis.speak(utter);
-        this.terminal.print('🔊 Speaking: "' + text.substring(0, 80) + (text.length > 80 ? '...' : '') + '"', 'info');
+        var voice = this._getBestVoice();
+        if (voice) utter.voice = voice;
+        utter.onstart = function() { self._updateButton(true); };
+        utter.onend = function() { self._updateButton(false); };
+        utter.onerror = function() { self._updateButton(false); };
+        // If voices not loaded yet, wait for voiceschanged with 2s timeout
+        if (this._voices.length === 0 && typeof speechSynthesis !== 'undefined') {
+            var spoken = false;
+            var onReady = function() {
+                if (spoken) return;
+                spoken = true;
+                var v = self._getBestVoice();
+                if (v) utter.voice = v;
+                speechSynthesis.cancel();
+                speechSynthesis.speak(utter);
+                speechSynthesis.onvoiceschanged = null;
+            };
+            speechSynthesis.onvoiceschanged = onReady;
+            setTimeout(function() {
+                if (!spoken) {
+                    spoken = true;
+                    speechSynthesis.onvoiceschanged = null;
+                    speechSynthesis.cancel();
+                    speechSynthesis.speak(utter);
+                }
+            }, 2000);
+        } else {
+            speechSynthesis.speak(utter);
+        }
     }
 
     speakText(text) {
-        // Called by AI assistant to auto-speak responses
         if (!this._autoSpeak || typeof speechSynthesis === 'undefined') return;
         speechSynthesis.cancel();
-        // Strip markdown/code blocks for cleaner speech
         const clean = text.replace(/```[\s\S]*?```/g, ' code block ').replace(/[*_#|`>]/g, '').replace(/\n+/g, '. ').trim();
         if (clean.length > 500) {
-            // Split long text into chunks
             const chunks = clean.match(/.{1,500}/g);
-            chunks.forEach(c => {
-                const u = new SpeechSynthesisUtterance(c);
-                u.rate = parseFloat(localStorage.getItem('webx11_tts_rate') || '1.0');
-                speechSynthesis.speak(u);
-            });
+            for (var i = 0; i < chunks.length; i++) {
+                this._doSpeak(chunks[i]);
+            }
         } else {
-            const u = new SpeechSynthesisUtterance(clean);
-            u.rate = parseFloat(localStorage.getItem('webx11_tts_rate') || '1.0');
-            speechSynthesis.speak(u);
+            this._doSpeak(clean);
         }
     }
     handle(args) {