feat(tts): Remove auto engine and fix ios STS

lib/core/services/settings_service.dart

@@ -9,10 +9,10 @@ import 'animation_service.dart';
 part 'settings_service.g.dart';
 
 /// Speech-to-text preference selection.
-enum SttPreference { auto, deviceOnly, serverOnly }
+enum SttPreference { deviceOnly, serverOnly }
 
 /// TTS engine selection
-enum TtsEngine { auto, device, server }
+enum TtsEngine { device, server }
 
 /// Service for managing app-wide settings including accessibility preferences
 class SettingsService {
@@ -232,15 +232,12 @@ class SettingsService {
 
   static TtsEngine _parseTtsEngine(String? raw) {
     switch ((raw ?? '').toLowerCase()) {
-      case 'auto':
-      case '':
-        return TtsEngine.auto;
       case 'server':
         return TtsEngine.server;
       case 'device':
         return TtsEngine.device;
       default:
-        return TtsEngine.auto;
+        return TtsEngine.device;
     }
   }
 
@@ -254,9 +251,8 @@ class SettingsService {
       case 'server_only':
       case 'server':
         return SttPreference.serverOnly;
-      case 'auto':
       default:
-        return SttPreference.auto;
+        return SttPreference.deviceOnly;
     }
   }
 
@@ -428,12 +424,12 @@ class AppSettings {
     this.socketTransportMode = 'ws',
     this.quickPills = const [],
     this.sendOnEnter = false,
-    this.sttPreference = SttPreference.auto,
+    this.sttPreference = SttPreference.deviceOnly,
     this.ttsVoice,
     this.ttsSpeechRate = 0.5,
     this.ttsPitch = 1.0,
     this.ttsVolume = 1.0,
-    this.ttsEngine = TtsEngine.auto,
+    this.ttsEngine = TtsEngine.device,
     this.ttsServerVoiceId,
     this.ttsServerVoiceName,
     this.voiceSilenceDuration = 2000,

lib/features/chat/services/text_to_speech_service.dart

@@ -23,7 +23,7 @@ class TextToSpeechService {
   final FlutterTts _tts = FlutterTts();
   final AudioPlayer _player = AudioPlayer();
   final ApiService? _api;
-  TtsEngine _engine = TtsEngine.auto;
+  TtsEngine _engine = TtsEngine.device;
   String? _preferredVoice;
   String? _serverPreferredVoice;
   double _speechRate = 0.5;
@@ -127,11 +127,9 @@ class TextToSpeechService {
     final serverAvailable = _api != null;
     switch (_engine) {
       case TtsEngine.device:
-        return _deviceEngineAvailable;
+        return _deviceEngineAvailable || serverAvailable;
       case TtsEngine.server:
         return serverAvailable;
-      case TtsEngine.auto:
-        return _deviceEngineAvailable || serverAvailable;
     }
   }
 
@@ -139,10 +137,7 @@ class TextToSpeechService {
     if (_engine == TtsEngine.server) {
       return _api != null;
     }
-    if (_engine == TtsEngine.device) {
+    // Device preference with graceful fallback to server if available.
-      return false;
-    }
-    // Auto: prefer device when available, otherwise fall back to server
     if (_deviceEngineAvailable) {
       return false;
     }
@@ -191,7 +186,7 @@ class TextToSpeechService {
     double speechRate = 0.5,
     double pitch = 1.0,
     double volume = 1.0,
-    TtsEngine engine = TtsEngine.auto,
+    TtsEngine engine = TtsEngine.device,
   }) async {
     if (_initialized) {
       _engine = engine;

lib/features/chat/services/voice_call_service.dart

@@ -133,9 +133,8 @@ class VoiceCallService {
     final hasLocalStt = _voiceInput.hasLocalStt;
     final hasServerStt = _voiceInput.hasServerStt;
     final ready = switch (_voiceInput.preference) {
-      SttPreference.deviceOnly => hasLocalStt,
+      SttPreference.deviceOnly => hasLocalStt || hasServerStt,
       SttPreference.serverOnly => hasServerStt,
-      SttPreference.auto => hasLocalStt || hasServerStt,
     };
 
     if (!ready) {
@@ -240,9 +239,8 @@ class VoiceCallService {
       final hasServerStt = _voiceInput.hasServerStt;
       final pref = _voiceInput.preference;
       final engineAvailable = switch (pref) {
-        SttPreference.deviceOnly => hasLocalStt,
+        SttPreference.deviceOnly => hasLocalStt || hasServerStt,
         SttPreference.serverOnly => hasServerStt,
-        SttPreference.auto => hasLocalStt || hasServerStt,
       };
 
       if (!engineAvailable) {

lib/features/chat/services/voice_input_service.dart

@@ -36,7 +36,7 @@ class VoiceInputService {
   bool _isInitialized = false;
   bool _isListening = false;
   bool _localSttAvailable = false;
-  SttPreference _preference = SttPreference.auto;
+  SttPreference _preference = SttPreference.deviceOnly;
   bool _usingServerStt = false;
   String? _selectedLocaleId;
   List<LocaleName> _locales = const [];
@@ -63,7 +63,6 @@ class VoiceInputService {
   bool get isSupportedPlatform => Platform.isAndroid || Platform.isIOS;
   bool get hasServerStt => _api != null;
   SttPreference get preference => _preference;
-  bool get allowsServerFallback => _preference != SttPreference.deviceOnly;
   bool get prefersServerOnly => _preference == SttPreference.serverOnly;
   bool get prefersDeviceOnly => _preference == SttPreference.deviceOnly;
 
@@ -101,15 +100,9 @@ class VoiceInputService {
       try {
         final sttGranted = await _speech.hasPermission();
         if (!sttGranted) {
-          if (prefersDeviceOnly) {
-            return false;
-          }
           _localSttAvailable = false;
         }
       } catch (_) {
-        if (prefersDeviceOnly) {
-          return false;
-        }
         _localSttAvailable = false;
       }
     }
@@ -248,11 +241,6 @@ class VoiceInputService {
           ? 'Speech recognition failed'
           : message,
     );
-    if (hasServerStt && allowsServerFallback) {
-      _textStreamController?.addError(exception);
-      unawaited(_beginServerFallback());
-      return;
-    }
     _textStreamController?.addError(exception);
     unawaited(_stopListening());
   }
@@ -265,6 +253,35 @@ class VoiceInputService {
     }
   }
 
+  Future<void> _startLocalRecognition({
+    required bool allowOnlineFallback,
+  }) async {
+    if (_selectedLocaleId != null) {
+      await _speech.setLanguage(_selectedLocaleId!);
+    }
+
+    Future<void> attempt(bool offline) => _speech.start(
+      SttRecognitionOptions(punctuation: true, offline: offline),
+    );
+
+    try {
+      await attempt(true);
+    } catch (error) {
+      if (Platform.isIOS && allowOnlineFallback) {
+        try {
+          await attempt(false);
+          return;
+        } catch (secondary) {
+          throw Exception(
+            'On-device speech failed ($error); '
+            'online fallback failed ($secondary).',
+          );
+        }
+      }
+      rethrow;
+    }
+  }
+
   Stream<String> startListening() {
     if (!_isInitialized) {
       throw Exception('Voice input not initialized');
@@ -304,11 +321,10 @@ class VoiceInputService {
           final isStillAvailable = await _speech.isSupported();
           if (!isStillAvailable && _isListening) {
             _localSttAvailable = false;
-            if (hasServerStt && allowsServerFallback) {
+            _textStreamController?.addError(
-              unawaited(_beginServerFallback());
+              Exception('On-device speech recognition unavailable'),
-            } else {
+            );
-              unawaited(_stopListening());
+            unawaited(_stopListening());
-            }
           }
         } catch (_) {
           // ignore availability check errors
@@ -338,19 +354,12 @@ class VoiceInputService {
 
       Future(() async {
         try {
-          if (_selectedLocaleId != null) {
+          await _startLocalRecognition(allowOnlineFallback: !prefersDeviceOnly);
-            await _speech.setLanguage(_selectedLocaleId!);
-          }
-          await _speech.start(SttRecognitionOptions(punctuation: true));
         } catch (error) {
           _localSttAvailable = false;
           if (!_isListening) return;
-          if (hasServerStt && allowsServerFallback) {
+          _textStreamController?.addError(error);
-            await _beginServerFallback();
+          await _stopListening();
-          } else {
-            _textStreamController?.addError(error);
-            await _stopListening();
-          }
         }
       });
     } else if (shouldUseServer) {
@@ -457,39 +466,6 @@ class VoiceInputService {
     }
   }
 
-  Future<void> _beginServerFallback() async {
-    if (!allowsServerFallback) {
-      _textStreamController?.addError(
-        Exception('Server speech-to-text disabled in preferences'),
-      );
-      await _stopListening();
-      return;
-    }
-    await _stopLocalStt();
-    if (!hasServerStt) {
-      _textStreamController?.addError(
-        Exception('Server speech-to-text unavailable'),
-      );
-      await _stopListening();
-      return;
-    }
-
-    _usingServerStt = true;
-    _autoStopTimer?.cancel();
-    _autoStopTimer = Timer(const Duration(seconds: 90), () {
-      if (_isListening) {
-        unawaited(_stopListening());
-      }
-    });
-
-    try {
-      await _startServerRecording();
-    } catch (error) {
-      _textStreamController?.addError(error);
-      await _stopListening();
-    }
-  }
-
   Future<void> _startServerRecording() async {
     await _setupVadStreams();
     final settings = _ref?.read(appSettingsProvider);
@@ -823,13 +799,11 @@ Future<bool> voiceInputAvailable(Ref ref) async {
   if (!initialized) return false;
   switch (service.preference) {
     case SttPreference.deviceOnly:
-      return service.hasLocalStt;
-    case SttPreference.serverOnly:
-      return service.hasServerStt;
-    case SttPreference.auto:
       if (service.hasLocalStt) return true;
       if (!service.hasServerStt) return false;
       break;
+    case SttPreference.serverOnly:
+      return service.hasServerStt;
   }
   final hasPermission = await service.checkPermissions();
   if (!hasPermission) return false;

lib/features/profile/views/app_customization_page.dart

@@ -501,8 +501,6 @@ class AppCustomizationPage extends ConsumerWidget {
       warnings.add(l10n.sttServerUnavailableWarning);
     }
 
-    final bool autoSelectable =
-        localAvailable || serverAvailable || localLoading;
     final bool deviceSelectable = localAvailable || localLoading;
     final bool serverSelectable = serverAvailable;
 
@@ -554,31 +552,6 @@ class AppCustomizationPage extends ConsumerWidget {
                 spacing: Spacing.sm,
                 runSpacing: Spacing.sm,
                 children: [
-                  ChoiceChip(
-                    label: Text(l10n.sttEngineAuto),
-                    selected: settings.sttPreference == SttPreference.auto,
-                    showCheckmark: false,
-                    selectedColor: theme.buttonPrimary,
-                    backgroundColor: theme.cardBackground,
-                    side: BorderSide(
-                      color: settings.sttPreference == SttPreference.auto
-                          ? theme.buttonPrimary.withValues(alpha: 0.6)
-                          : theme.textPrimary.withValues(alpha: 0.2),
-                    ),
-                    labelStyle: TextStyle(
-                      color: settings.sttPreference == SttPreference.auto
-                          ? theme.buttonPrimaryText
-                          : theme.textPrimary,
-                      fontWeight: FontWeight.w600,
-                    ),
-                    onSelected: autoSelectable
-                        ? (value) {
-                            if (value) {
-                              notifier.setSttPreference(SttPreference.auto);
-                            }
-                          }
-                        : null,
-                  ),
                   ChoiceChip(
                     label: Text(l10n.sttEngineDevice),
                     selected:
@@ -684,9 +657,7 @@ class AppCustomizationPage extends ConsumerWidget {
                   ),
                 ),
               ],
-              if (settings.sttPreference == SttPreference.serverOnly ||
+              if (settings.sttPreference == SttPreference.serverOnly) ...[
-                  (settings.sttPreference == SttPreference.auto &&
-                      serverAvailable)) ...[
                 const SizedBox(height: Spacing.md),
                 const Divider(),
                 const SizedBox(height: Spacing.md),
@@ -785,20 +756,11 @@ class AppCustomizationPage extends ConsumerWidget {
     final bool deviceAvailable =
         ttsService.deviceEngineAvailable || !ttsService.isInitialized;
     final bool serverAvailable = ttsService.serverEngineAvailable;
-    final bool autoSelectable = deviceAvailable || serverAvailable;
     final bool deviceSelectable = deviceAvailable;
     final bool serverSelectable = serverAvailable;
     final ttsDescription = _ttsPreferenceDescription(l10n, settings);
     final warnings = <String>[];
     switch (settings.ttsEngine) {
-      case TtsEngine.auto:
-        if (!deviceAvailable) {
-          warnings.add(l10n.ttsDeviceUnavailableWarning);
-        }
-        if (!serverAvailable) {
-          warnings.add(l10n.ttsServerUnavailableWarning);
-        }
-        break;
       case TtsEngine.device:
         if (!deviceAvailable) {
           warnings.add(l10n.ttsDeviceUnavailableWarning);
@@ -852,37 +814,6 @@ class AppCustomizationPage extends ConsumerWidget {
                 spacing: Spacing.sm,
                 runSpacing: Spacing.sm,
                 children: [
-                  ChoiceChip(
-                    label: Text(l10n.ttsEngineAuto),
-                    selected: settings.ttsEngine == TtsEngine.auto,
-                    showCheckmark: false,
-                    selectedColor: theme.buttonPrimary,
-                    backgroundColor: theme.cardBackground,
-                    side: BorderSide(
-                      color: settings.ttsEngine == TtsEngine.auto
-                          ? theme.buttonPrimary.withValues(alpha: 0.6)
-                          : theme.textPrimary.withValues(
-                              alpha: autoSelectable ? 0.2 : 0.12,
-                            ),
-                    ),
-                    labelStyle: TextStyle(
-                      color: settings.ttsEngine == TtsEngine.auto
-                          ? theme.buttonPrimaryText
-                          : theme.textPrimary.withValues(
-                              alpha: autoSelectable ? 1.0 : 0.45,
-                            ),
-                      fontWeight: FontWeight.w600,
-                    ),
-                    onSelected: autoSelectable
-                        ? (value) {
-                            if (value) {
-                              ref
-                                  .read(appSettingsProvider.notifier)
-                                  .setTtsEngine(TtsEngine.auto);
-                            }
-                          }
-                        : null,
-                  ),
                   ChoiceChip(
                     label: Text(l10n.ttsEngineDevice),
                     selected: settings.ttsEngine == TtsEngine.device,
@@ -1060,8 +991,6 @@ class AppCustomizationPage extends ConsumerWidget {
     SttPreference preference,
   ) {
     switch (preference) {
-      case SttPreference.auto:
-        return l10n.sttEngineAutoDescription;
       case SttPreference.deviceOnly:
         return l10n.sttEngineDeviceDescription;
       case SttPreference.serverOnly:
@@ -1074,8 +1003,6 @@ class AppCustomizationPage extends ConsumerWidget {
     AppSettings settings,
   ) {
     switch (settings.ttsEngine) {
-      case TtsEngine.auto:
-        return l10n.ttsEngineAutoDescription;
       case TtsEngine.device:
         return l10n.ttsEngineDeviceDescription;
       case TtsEngine.server:
@@ -1093,8 +1020,6 @@ class AppCustomizationPage extends ConsumerWidget {
     final serverName = _getDisplayVoiceName(serverVoice, l10n.ttsSystemDefault);
 
     switch (settings.ttsEngine) {
-      case TtsEngine.auto:
-        return '${l10n.ttsEngineDevice}: $deviceName • ${l10n.ttsEngineServer}: $serverName';
       case TtsEngine.device:
         return deviceName;
       case TtsEngine.server:

lib/l10n/app_de.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Stille-Dauer",
   "sttSilenceDurationDescription": "Zeit nach Stille warten, bevor die Aufnahme automatisch gestoppt wird",
   "ttsSettings": "Text zu Sprache",
-  "ttsEngineAuto": "Automatisch",
-  "ttsEngineAutoDescription": "Verwendet die Sprachausgabe auf dem Gerät, wenn verfügbar, und greift sonst auf deinen Server zurück.",
   "ttsEngineDeviceDescription": "Behält die Ausgabe auf diesem Gerät. Sprachausgabe funktioniert nicht, wenn das Gerät keine TTS-Unterstützung bietet.",
   "ttsEngineServerDescription": "Sendet die Ausgabe immer an deinen OpenWebUI-Server.",
   "ttsDeviceUnavailableWarning": "Sprachausgabe auf dem Gerät steht auf diesem Gerät nicht zur Verfügung.",

lib/l10n/app_en.arb

@@ -1275,10 +1275,6 @@
   "@ttsEngineLabel": {
     "description": "Label for selecting the text-to-speech engine."
   },
-  "ttsEngineAuto": "Auto",
-  "@ttsEngineAuto": {
-    "description": "Chip label for automatically selecting the text-to-speech engine."
-  },
   "ttsEngineDevice": "On device",
   "@ttsEngineDevice": {
     "description": "Chip label for using on-device text-to-speech."
@@ -1287,10 +1283,6 @@
   "@ttsEngineServer": {
     "description": "Chip label for using server-side text-to-speech."
   },
-  "ttsEngineAutoDescription": "Use on-device speech when available and fall back to your server.",
-  "@ttsEngineAutoDescription": {
-    "description": "Description shown when automatic text-to-speech preference is active."
-  },
   "ttsEngineDeviceDescription": "Keep synthesis on this device. Voice playback stops working if on-device TTS isn’t supported.",
   "@ttsEngineDeviceDescription": {
     "description": "Description shown when on-device text-to-speech preference is active."

lib/l10n/app_es.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Duración del silencio",
   "sttSilenceDurationDescription": "Tiempo de espera después del silencio antes de detener automáticamente la grabación",
   "ttsSettings": "Texto a voz",
-  "ttsEngineAuto": "Automático",
-  "ttsEngineAutoDescription": "Usa la síntesis en el dispositivo cuando esté disponible y, si no, recurre a tu servidor.",
   "ttsEngineDeviceDescription": "Mantiene la síntesis en este dispositivo. La reproducción de voz no funciona si el dispositivo no admite TTS.",
   "ttsEngineServerDescription": "Solicita siempre el audio a tu servidor OpenWebUI.",
   "ttsDeviceUnavailableWarning": "La síntesis de voz en el dispositivo no está disponible en este dispositivo.",

lib/l10n/app_fr.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Durée du silence",
   "sttSilenceDurationDescription": "Temps d'attente après le silence avant d'arrêter automatiquement l'enregistrement",
   "ttsSettings": "Synthèse vocale",
-  "ttsEngineAuto": "Auto",
-  "ttsEngineAutoDescription": "Utilise la synthèse locale quand c’est possible, sinon bascule vers votre serveur.",
   "ttsEngineDeviceDescription": "Garde la synthèse sur cet appareil. La lecture vocale ne fonctionne plus si l’appareil n’offre pas la synthèse vocale.",
   "ttsEngineServerDescription": "Demande toujours l'audio à votre serveur OpenWebUI.",
   "ttsDeviceUnavailableWarning": "La synthèse vocale sur l’appareil n’est pas disponible sur cet appareil.",

lib/l10n/app_it.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Durata del silenzio",
   "sttSilenceDurationDescription": "Tempo di attesa dopo il silenzio prima di fermare automaticamente la registrazione",
   "ttsSettings": "Sintesi vocale",
-  "ttsEngineAuto": "Automatico",
-  "ttsEngineAutoDescription": "Usa la sintesi sul dispositivo quando disponibile e altrimenti passa al tuo server.",
   "ttsEngineDeviceDescription": "Mantiene la sintesi su questo dispositivo. La riproduzione vocale non funziona se il dispositivo non supporta il TTS.",
   "ttsEngineServerDescription": "Richiede sempre l'audio dal tuo server OpenWebUI.",
   "ttsDeviceUnavailableWarning": "La sintesi vocale sul dispositivo non è disponibile su questo dispositivo.",

lib/l10n/app_ko.arb

@@ -447,10 +447,8 @@
   "sttSilenceDuration": "침묵 지속 시간",
   "sttSilenceDurationDescription": "침묵 후 자동 녹음 중지 전 대기 시간",
   "ttsEngineLabel": "엔진",
-  "ttsEngineAuto": "자동",
   "ttsEngineDevice": "기기에서",
   "ttsEngineServer": "서버",
-  "ttsEngineAutoDescription": "사용 가능한 경우 기기에서 음성을 사용하고 서버로 폴백합니다.",
   "ttsEngineDeviceDescription": "합성을 이 기기에서 유지합니다. 기기에서 TTS가 지원되지 않으면 음성 재생이 작동하지 않습니다.",
   "ttsEngineServerDescription": "항상 OpenWebUI 서버에서 오디오를 요청합니다.",
   "ttsDeviceUnavailableWarning": "이 기기에서 기기 텍스트 음성 변환을 사용할 수 없습니다.",

lib/l10n/app_nl.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Stilteduur",
   "sttSilenceDurationDescription": "Tijd om te wachten na stilte voordat de opname automatisch stopt",
   "ttsSettings": "Tekst naar spraak",
-  "ttsEngineAuto": "Automatisch",
-  "ttsEngineAutoDescription": "Gebruikt spraaksynthese op het apparaat wanneer beschikbaar en valt anders terug op je server.",
   "ttsEngineDeviceDescription": "Houdt de synthese op dit apparaat. Spraakweergave werkt niet als het apparaat geen TTS ondersteunt.",
   "ttsEngineServerDescription": "Vraagt altijd audio op bij je OpenWebUI-server.",
   "ttsDeviceUnavailableWarning": "Spraaksynthese op het apparaat is niet beschikbaar op dit apparaat.",

lib/l10n/app_ru.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "Длительность тишины",
   "sttSilenceDurationDescription": "Время ожидания после тишины перед автоматической остановкой записи",
   "ttsSettings": "Преобразование текста в речь",
-  "ttsEngineAuto": "Авто",
-  "ttsEngineAutoDescription": "Использует синтез речи на устройстве, когда это возможно, иначе переключается на ваш сервер.",
   "ttsEngineDeviceDescription": "Оставляет синтез на этом устройстве. Воспроизведение голоса не работает, если устройство не поддерживает синтез речи.",
   "ttsEngineServerDescription": "Всегда запрашивает аудио у вашего сервера OpenWebUI.",
   "ttsDeviceUnavailableWarning": "Синтез речи на устройстве недоступен на этом устройстве.",

lib/l10n/app_zh.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "静音持续时间",
   "sttSilenceDurationDescription": "检测到静音后等待多久自动停止录音",
   "ttsSettings": "文本转语音",
-  "ttsEngineAuto": "自动",
-  "ttsEngineAutoDescription": "在可用时使用本机合成，否则切换到你的服务器。",
   "ttsEngineDeviceDescription": "在此设备上完成合成。如果设备不支持文本转语音，语音播放将不可用。",
   "ttsEngineServerDescription": "始终向你的 OpenWebUI 服务器请求音频。",
   "ttsDeviceUnavailableWarning": "此设备不支持本机文本转语音。",

lib/l10n/app_zh_Hant.arb

@@ -321,8 +321,6 @@
   "sttSilenceDuration": "靜音持續時間",
   "sttSilenceDurationDescription": "檢測到靜音後等待多久自動停止錄音",
   "ttsSettings": "文本轉語音",
-  "ttsEngineAuto": "自動",
-  "ttsEngineAutoDescription": "在可用時使用本機合成，否則切換到你的服務器。",
   "ttsEngineDeviceDescription": "在此設備上完成合成。如果設備不支持文本轉語音，語音播放將不可用。",
   "ttsEngineServerDescription": "始終向你的 OpenWebUI 服務器請求音頻。",
   "ttsDeviceUnavailableWarning": "此設備不支持本機文本轉語音。",