refactor

apps/german-app/src/App.vue

@@ -73,7 +73,7 @@ export default {
           { type: 'wg_viewing', name: 'WG Room Viewing', emoji: '🏠' },
           { type: 'burgeramt', name: 'At the Bürgeramt', emoji: '🏛️' },
           { type: 'biergarten', name: 'At a Biergarten', emoji: '🍺' },
-          { type: 'ubahn', name: 'U-Bahn Help', emoji: '🚇' }
+          { type: 'ber_airport', name: 'BER Airport Train Help', emoji: '✈️' }
         ]
       }
     },
@@ -83,7 +83,7 @@ export default {
         'wg_viewing': '🏠',
         'burgeramt': '🏛️',
         'biergarten': '🍺',
-        'ubahn': '🚇'
+        'ber_airport': '✈️'
       }
       return emojiMap[type] || '📍'
     },

apps/german-app/src/components/GermanSpeechInterface.vue

@@ -1,5 +1,41 @@
 <template>
   <div class="speech-interface">
+    <!-- Tutorial Popup -->
+    <div v-if="showTutorial" class="tutorial-overlay" @click="closeTutorial">
+      <div class="tutorial-popup" @click.stop>
+        <div class="tutorial-header">
+          <h3>Willkommen bei Street Lingo! 🎉</h3>
+          <button @click="closeTutorial" class="tutorial-close">✕</button>
+        </div>
+        <div class="tutorial-content">
+          <div class="tutorial-step">
+            <div class="tutorial-icon">🎙️</div>
+            <div class="tutorial-text">
+              <strong>Sprechen üben</strong>
+              <p>Klicken Sie auf "Gespräch beginnen", um mit KI-Charakteren in realen deutschen Szenarien zu sprechen</p>
+            </div>
+          </div>
+          <div class="tutorial-step">
+            <div class="tutorial-icon">💡</div>
+            <div class="tutorial-text">
+              <strong>Hilfe erhalten</strong>
+              <p>Aktivieren Sie die "Hilfe"-Schaltfläche und pausieren Sie während des Gesprächs, um hilfreiche Phrasen zu erhalten</p>
+            </div>
+          </div>
+          <div class="tutorial-step">
+            <div class="tutorial-icon">📊</div>
+            <div class="tutorial-text">
+              <strong>Gesprächsanalyse</strong>
+              <p>Klicken Sie auf "Gespräch beenden", um personalisiertes Feedback zu Ihrem Deutsch zu erhalten</p>
+            </div>
+          </div>
+        </div>
+        <div class="tutorial-footer">
+          <button @click="closeTutorial" class="tutorial-got-it">Verstanden!</button>
+        </div>
+      </div>
+    </div>
+    
     <div class="conversation-area">
       <div class="messages" ref="messagesContainer">
         <div 
@@ -41,11 +77,6 @@
         <span class="transcription-status">{{ getTranscriptionStatus() }}</span>
       </div>
       
-      <!-- Recording Stopped Notification -->
-      <div v-if="showRecordingStoppedNotification" class="recording-stopped-notification">
-        <p>🛑 Aufnahme gestoppt. Klicken Sie auf "Sprechen", um fortzufahren.</p>
-      </div>
-      
       <!-- Inline Suggestion Panel -->
       <div v-if="showSuggestionPopup" class="suggestion-panel">
         <div class="suggestion-panel-header">
@@ -73,7 +104,18 @@
     </div>
 
     <div class="controls">
+      <!-- Start Conversation Button -->
       <button 
+        v-if="showStartButton"
+        @click="startConversation"
+        class="start-conversation-btn"
+        :disabled="isConnecting"
+      >
+        🎙️ Gespräch beginnen
+      </button>
+      
+      <button 
+        v-else
         @click="toggleRecording"
         :class="['record-btn', { recording: isRecording }]"
         :disabled="isConnecting || isFinished"
@@ -81,16 +123,6 @@
         {{ isRecording ? '🛑 Stopp' : '🎤 Sprechen' }}
       </button>
       
-      <button 
-        v-if="isRecording"
-        @click="forceStopRecording"
-        class="force-stop-btn"
-        title="Sofort stoppen"
-        :disabled="isFinished"
-      >
-        ⏹️
-      </button>
-      
       <div class="suggestion-toggle-container">
         <label class="suggestion-toggle-label" for="suggestion-toggle">
           💡 Hilfe
@@ -131,9 +163,9 @@
         @click="finishConversation"
         class="finish-btn"
         title="Gespräch beenden und Bewertung erhalten"
-        :disabled="messages.length === 0 || isFinished"
+        :disabled="messages.length === 0 || isFinished || isLoadingFeedback"
       >
-        ✓ Gespräch beenden
+        {{ isLoadingFeedback ? 'Feedback wird analysiert...' : '✓ Gespräch beenden' }}
       </button>
     </div>
 
@@ -255,12 +287,21 @@ export default {
       isAICurrentlySpeaking: false,
       lastAIResponseTime: null,
       isFinished: false,
-      conversationFeedback: null
+      conversationFeedback: null,
+      isLoadingFeedback: false,
+      hasRequestedInitialGreeting: false,
+      showStartButton: true,
+      showTutorial: false
     }
   },
   mounted() {
     this.connectWebSocket()
     this.loadScenarioData()
+    
+    // Check if user is new and show tutorial
+    this.checkAndShowTutorial()
+    
+    // Don't automatically request greeting on mount - wait for user to click start button
   },
   beforeUnmount() {
     this.disconnect()
@@ -272,11 +313,8 @@ export default {
         if (oldScenario && newScenario !== oldScenario) {
           this.resetConversationOnScenarioChange()
         } else if (newScenario && !oldScenario) {
-          // Initial scenario load - request greeting if connected
+          // Initial scenario load - don't auto request greeting, wait for user to click start button
           console.log('Scenario initially set to:', newScenario)
-          if (this.connectionStatus === 'connected') {
-            this.requestInitialGreeting()
-          }
         }
       }
     },
@@ -300,8 +338,7 @@ export default {
           this.isConnecting = false
           console.log('German WebSocket connected')
           
-          // Request initial greeting from character
+          // Don't automatically request greeting - wait for user to click start button
-          this.requestInitialGreeting()
         }
         
         this.websocket.onmessage = (event) => {
@@ -617,16 +654,18 @@ export default {
       this.conversationComplete = false
       this.isFinished = false
       this.conversationFeedback = null
+      this.hasRequestedInitialGreeting = false
       
       if (this.websocket && this.websocket.readyState === WebSocket.OPEN) {
         const resetMessage = {
           type: 'conversation_reset'
         }
         this.websocket.send(JSON.stringify(resetMessage))
-        
-        // Request initial greeting for new scenario
-        this.requestInitialGreeting()
       }
+      
+      // For scenario changes, show start button again so user must start new conversation
+      this.showStartButton = true
+      // Don't automatically request greeting - wait for user to click start button
     },
 
     async requestTranslation(message) {
@@ -666,7 +705,7 @@ export default {
     },
 
     requestInitialGreeting() {
-      if (this.websocket && this.websocket.readyState === WebSocket.OPEN) {
+      if (this.websocket && this.websocket.readyState === WebSocket.OPEN && !this.hasRequestedInitialGreeting) {
         // Add a small delay to ensure scenario is properly set
         setTimeout(() => {
           if (this.scenario) {
@@ -675,16 +714,47 @@ export default {
               scenario_context: this.scenario
             }
             this.websocket.send(JSON.stringify(greetingMessage))
+            this.hasRequestedInitialGreeting = true
             console.log('Sent initial greeting for scenario:', this.scenario)
           } else {
-            console.log('No scenario set, retrying in 200ms')
+            console.log('No scenario set, retrying in 300ms')
-            // Retry if scenario not set yet
+            // Retry if scenario not set yet with longer delay
-            setTimeout(() => this.requestInitialGreeting(), 200)
+            setTimeout(() => this.requestInitialGreeting(), 300)
           }
-        }, 100)
+        }, 200)  // Increased delay to ensure scenario is set
       }
     },
 
+    // Helper method to check if we should request greeting
+    checkAndRequestGreeting() {
+      if (this.websocket && this.websocket.readyState === WebSocket.OPEN && 
+          this.scenario && !this.hasRequestedInitialGreeting) {
+        this.requestInitialGreeting()
+      }
+    },
+
+    startConversation() {
+      // Hide the start button and show normal controls
+      this.showStartButton = false
+      
+      // Request the initial greeting now that user has interacted
+      this.checkAndRequestGreeting()
+    },
+
+    checkAndShowTutorial() {
+      // Check if user has seen tutorial before
+      const hasSeenTutorial = localStorage.getItem('streetLingo_tutorialSeen')
+      if (!hasSeenTutorial) {
+        this.showTutorial = true
+      }
+    },
+
+    closeTutorial() {
+      this.showTutorial = false
+      // Mark tutorial as seen
+      localStorage.setItem('streetLingo_tutorialSeen', 'true')
+    },
+
     setupAutoRecording(message) {
       // Wait for the audio element to be created in the DOM
       this.$nextTick(() => {
@@ -939,6 +1009,7 @@ export default {
       
       // Mark conversation as finished
       this.isFinished = true
+      this.isLoadingFeedback = true
       
       // Close suggestion popup if open
       if (this.showSuggestionPopup) {
@@ -966,6 +1037,7 @@ export default {
         if (response.ok) {
           const feedback = await response.json()
           this.conversationFeedback = feedback
+          this.isLoadingFeedback = false
           
           // Scroll to show feedback
           this.$nextTick(() => {
@@ -979,6 +1051,7 @@ export default {
             suggestions: [],
             examples: []
           }
+          this.isLoadingFeedback = false
         }
       } catch (error) {
         console.error('Error getting conversation feedback:', error)
@@ -988,6 +1061,7 @@ export default {
           suggestions: [],
           examples: []
         }
+        this.isLoadingFeedback = false
       }
     },
 
@@ -1805,4 +1879,203 @@ export default {
   transform: none;
   box-shadow: none;
 }
+
+/* Start Conversation Button Styles */
+.start-conversation-btn {
+  background: var(--primary);
+  color: white;
+  border: none;
+  padding: 1rem 2rem;
+  border-radius: var(--radius-lg);
+  cursor: pointer;
+  font-family: 'DM Sans', sans-serif;
+  font-size: 0.9rem;
+  font-weight: 500;
+  transition: all 0.3s cubic-bezier(0.4, 0, 0.2, 1);
+  box-shadow: var(--shadow);
+  display: flex;
+  align-items: center;
+  gap: 0.5rem;
+  min-width: 200px;
+  justify-content: center;
+}
+
+.start-conversation-btn:hover:not(:disabled) {
+  background: var(--primary-dark);
+  transform: translateY(-2px);
+  box-shadow: var(--shadow-lg);
+}
+
+.start-conversation-btn:disabled {
+  background: var(--text-muted);
+  cursor: not-allowed;
+  transform: none;
+  box-shadow: var(--shadow-sm);
+}
+
+/* Tutorial Popup Styles */
+.tutorial-overlay {
+  position: fixed;
+  top: 0;
+  left: 0;
+  right: 0;
+  bottom: 0;
+  background: rgba(0, 0, 0, 0.6);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  z-index: 9999;
+  backdrop-filter: blur(4px);
+}
+
+.tutorial-popup {
+  background: var(--surface);
+  border-radius: var(--radius-lg);
+  box-shadow: var(--shadow-lg);
+  max-width: 500px;
+  width: 90%;
+  max-height: 80vh;
+  overflow-y: auto;
+  animation: tutorialSlideIn 0.3s ease-out;
+}
+
+@keyframes tutorialSlideIn {
+  from {
+    opacity: 0;
+    transform: translateY(-20px) scale(0.95);
+  }
+  to {
+    opacity: 1;
+    transform: translateY(0) scale(1);
+  }
+}
+
+.tutorial-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  padding: 1.5rem 1.5rem 1rem 1.5rem;
+  border-bottom: 1px solid var(--border);
+}
+
+.tutorial-header h3 {
+  margin: 0;
+  color: var(--text);
+  font-size: 1.3rem;
+  font-weight: 700;
+}
+
+.tutorial-close {
+  background: none;
+  border: none;
+  font-size: 1.2rem;
+  cursor: pointer;
+  color: var(--text-muted);
+  padding: 0.25rem;
+  border-radius: var(--radius);
+  transition: all 0.2s ease;
+}
+
+.tutorial-close:hover {
+  background: var(--surface-alt);
+  color: var(--text);
+}
+
+.tutorial-content {
+  padding: 1.5rem;
+}
+
+.tutorial-step {
+  display: flex;
+  align-items: flex-start;
+  gap: 1rem;
+  margin-bottom: 1.5rem;
+}
+
+.tutorial-step:last-child {
+  margin-bottom: 0;
+}
+
+.tutorial-icon {
+  font-size: 2rem;
+  flex-shrink: 0;
+}
+
+.tutorial-text {
+  flex: 1;
+}
+
+.tutorial-text strong {
+  display: block;
+  color: var(--text);
+  font-size: 1rem;
+  font-weight: 600;
+  margin-bottom: 0.5rem;
+}
+
+.tutorial-text p {
+  margin: 0;
+  color: var(--text-light);
+  font-size: 0.9rem;
+  line-height: 1.4;
+}
+
+.tutorial-footer {
+  padding: 1rem 1.5rem 1.5rem 1.5rem;
+  border-top: 1px solid var(--border);
+  display: flex;
+  justify-content: center;
+}
+
+.tutorial-got-it {
+  background: var(--primary);
+  color: white;
+  border: none;
+  padding: 0.75rem 2rem;
+  border-radius: var(--radius-lg);
+  cursor: pointer;
+  font-family: 'DM Sans', sans-serif;
+  font-size: 0.9rem;
+  font-weight: 600;
+  transition: all 0.2s ease;
+  box-shadow: var(--shadow-sm);
+}
+
+.tutorial-got-it:hover {
+  background: var(--primary-dark);
+  transform: translateY(-1px);
+  box-shadow: var(--shadow);
+}
+
+@media (max-width: 768px) {
+  .tutorial-popup {
+    width: 95%;
+    max-height: 85vh;
+  }
+  
+  .tutorial-header {
+    padding: 1rem 1rem 0.75rem 1rem;
+  }
+  
+  .tutorial-header h3 {
+    font-size: 1.1rem;
+  }
+  
+  .tutorial-content {
+    padding: 1rem;
+  }
+  
+  .tutorial-step {
+    gap: 0.75rem;
+    margin-bottom: 1rem;
+  }
+  
+  .tutorial-icon {
+    font-size: 1.5rem;
+  }
+  
+  .tutorial-footer {
+    padding: 0.75rem 1rem 1rem 1rem;
+  }
+}
 </style>

apps/german-app/src/views/HomeView.vue

@@ -143,11 +143,11 @@ export default {
             goal: 'Order a beer and traditional German food'
           },
           {
-            type: 'ubahn',
+            type: 'ber_airport',
-            name: 'U-Bahn Help',
+            name: 'BER Airport Train Help',
-            description: 'Get help with public transport in Berlin',
+            description: 'Get train directions from BER airport to Potsdamer Platz',
-            challenge: 'Transport terminology and directions',
+            challenge: 'Airport-to-city transport and dealing with delays',
-            goal: 'Get directions and buy appropriate ticket'
+            goal: 'Get clear directions to Potsdamer Platz despite train delays'
           }
         ]
       }
@@ -158,7 +158,8 @@ export default {
         'wg_viewing': '🏠',
         'burgeramt': '🏛️',
         'biergarten': '🍺',
-        'ubahn': '🚇'
+        'ber_airport': '✈️',
+        'arzt': '👨‍⚕️'
       }
       return emojiMap[type] || '📍'
     },

apps/german-app/src/views/ScenarioView.vue

@@ -114,7 +114,8 @@ export default {
         'wg_viewing': '🏠',
         'burgeramt': '🏛️',
         'biergarten': '🍺',
-        'ubahn': '🚇'
+        'ber_airport': '✈️',
+        'arzt': '👨‍⚕️'
       }
       return emojiMap[type] || '📍'
     },
@@ -124,7 +125,8 @@ export default {
         'wg_viewing': '👩‍🎓',
         'burgeramt': '👩‍💼',
         'biergarten': '👨‍🍳',
-        'ubahn': '👨‍🚀'
+        'ber_airport': '👩‍💼',
+        'arzt': '👨‍⚕️'
       }
       return avatarMap[type] || '👤'
     }

apps/indonesian-app/index.html

@@ -3,7 +3,7 @@
 <head>
     <meta charset="UTF-8">
     <meta name="viewport" content="width=device-width, initial-scale=1.0">
-    <title>Learn Indonesian - Realistic Scenarios</title>
+    <title>Street Lingo - Learn Indonesian through everyday scenarios</title>
 </head>
 <body>
     <div id="app"></div>

apps/indonesian-app/src/App.vue

@@ -5,7 +5,7 @@
         <div class="logo-section">
           <div class="logo">
             <span class="flag">🇮🇩</span>
-            <h1>Learn Indonesian</h1>
+            <h1>Street Lingo</h1>
           </div>
           <p class="tagline">Learn Indonesian through everyday scenarios</p>
         </div>

apps/indonesian-app/src/components/SpeechInterface.vue

@@ -1,5 +1,41 @@
 <template>
   <div class="speech-interface">
+    <!-- Tutorial Popup -->
+    <div v-if="showTutorial" class="tutorial-overlay" @click="closeTutorial">
+      <div class="tutorial-popup" @click.stop>
+        <div class="tutorial-header">
+          <h3>Welcome to Street Lingo! 🎉</h3>
+          <button @click="closeTutorial" class="tutorial-close">✕</button>
+        </div>
+        <div class="tutorial-content">
+          <div class="tutorial-step">
+            <div class="tutorial-icon">🎙️</div>
+            <div class="tutorial-text">
+              <strong>Practice Speaking</strong>
+              <p>Click "Start Conversation" to begin chatting with AI characters in real Indonesian scenarios</p>
+            </div>
+          </div>
+          <div class="tutorial-step">
+            <div class="tutorial-icon">💡</div>
+            <div class="tutorial-text">
+              <strong>Get Help</strong>
+              <p>Enable the "Help" toggle and pause during conversation to get helpful phrase suggestions</p>
+            </div>
+          </div>
+          <div class="tutorial-step">
+            <div class="tutorial-icon">📊</div>
+            <div class="tutorial-text">
+              <strong>Conversation Analysis</strong>
+              <p>Click "Finish Conversation" to get personalized feedback on your Indonesian</p>
+            </div>
+          </div>
+        </div>
+        <div class="tutorial-footer">
+          <button @click="closeTutorial" class="tutorial-got-it">Got it!</button>
+        </div>
+      </div>
+    </div>
+    
     <div class="conversation-area">
       <div class="messages" ref="messagesContainer">
         <div 
@@ -73,7 +109,18 @@
     </div>
 
     <div class="controls">
+      <!-- Start Conversation Button -->
       <button 
+        v-if="showStartButton"
+        @click="startConversation"
+        class="start-conversation-btn"
+        :disabled="isConnecting"
+      >
+        🎙️ Start Conversation
+      </button>
+      
+      <button 
+        v-else
         @click="toggleRecording"
         :class="['record-btn', { recording: isRecording }]"
         :disabled="isConnecting || isFinished"
@@ -81,15 +128,6 @@
         {{ isRecording ? '🛑 Stop' : '🎤 Speak' }}
       </button>
       
-      <button 
-        v-if="isRecording"
-        @click="forceStopRecording"
-        class="force-stop-btn"
-        title="Force stop"
-        :disabled="isFinished"
-      >
-        ⏹️
-      </button>
       
       <div class="suggestion-toggle-container">
         <label class="suggestion-toggle-label" for="suggestion-toggle">
@@ -158,6 +196,14 @@
       </div>
     </div>
 
+    <!-- Feedback Loading Notification -->
+    <div class="feedback-loading" v-if="isLoadingFeedback">
+      <div class="loading-content">
+        <div class="loading-spinner"></div>
+        <p>Analyzing your conversation...</p>
+      </div>
+    </div>
+
     <!-- Conversation Feedback -->
     <div class="feedback-section" v-if="conversationFeedback">
       <h3 class="feedback-title">
@@ -254,12 +300,21 @@ export default {
       showRecordingStoppedNotification: false,
       lastAIResponseTime: null,
       isFinished: false,
-      conversationFeedback: null
+      conversationFeedback: null,
+      isLoadingFeedback: false,
+      hasRequestedInitialGreeting: false,
+      showStartButton: true,
+      showTutorial: false
     }
   },
   mounted() {
     this.connectWebSocket()
     this.loadScenarioData()
+    
+    // Check if user is new and show tutorial
+    this.checkAndShowTutorial()
+    
+    // Don't automatically request greeting on mount - wait for user to click start button
   },
   beforeUnmount() {
     this.disconnect()
@@ -272,11 +327,8 @@ export default {
           // Reset conversation when scenario changes
           this.resetConversationOnScenarioChange()
         } else if (newScenario && !oldScenario) {
-          // Initial scenario load - request greeting if connected
+          // Initial scenario load - don't auto request greeting, wait for user to click start button
           console.log('Scenario initially set to:', newScenario)
-          if (this.connectionStatus === 'connected') {
-            this.requestInitialGreeting()
-          }
         }
       }
     },
@@ -301,8 +353,7 @@ export default {
           this.isConnecting = false
           console.log('WebSocket connected')
           
-          // Request initial greeting from character
+          // Don't automatically request greeting - wait for user to click start button
-          this.requestInitialGreeting()
         }
         
         this.websocket.onmessage = (event) => {
@@ -459,7 +510,7 @@ export default {
       
       this.isRecording = false
       this.isAutoListening = false
-      this.currentTranscription = 'Processing...'
+      this.currentTranscription = ''  // Don't show "Processing..." delay
       this.isTranscriptionFinal = false
     },
     
@@ -558,6 +609,8 @@ export default {
       this.isTranscriptionFinal = false
       this.isFinished = false
       this.conversationFeedback = null
+      this.isLoadingFeedback = false
+      this.hasRequestedInitialGreeting = false
       
       // Reset goals
       this.resetGoals()
@@ -614,6 +667,8 @@ export default {
       this.conversationComplete = false
       this.isFinished = false
       this.conversationFeedback = null
+      this.isLoadingFeedback = false
+      this.hasRequestedInitialGreeting = false
     },
 
     resetConversationOnScenarioChange() {
@@ -627,6 +682,8 @@ export default {
       this.conversationComplete = false
       this.isFinished = false
       this.conversationFeedback = null
+      this.isLoadingFeedback = false
+      this.hasRequestedInitialGreeting = false
       
       // Send reset message to backend if connected
       if (this.websocket && this.websocket.readyState === WebSocket.OPEN) {
@@ -634,10 +691,11 @@ export default {
           type: 'conversation_reset'
         }
         this.websocket.send(JSON.stringify(resetMessage))
-        
-        // Request initial greeting for new scenario
-        this.requestInitialGreeting()
       }
+      
+      // For scenario changes, show start button again so user must start new conversation
+      this.showStartButton = true
+      // Don't automatically request greeting - wait for user to click start button
     },
 
     async requestTranslation(message) {
@@ -678,7 +736,7 @@ export default {
     },
 
     requestInitialGreeting() {
-      if (this.websocket && this.websocket.readyState === WebSocket.OPEN) {
+      if (this.websocket && this.websocket.readyState === WebSocket.OPEN && !this.hasRequestedInitialGreeting) {
         // Add a small delay to ensure scenario is properly set
         setTimeout(() => {
           if (this.scenario) {
@@ -687,16 +745,48 @@ export default {
               scenario_context: this.scenario
             }
             this.websocket.send(JSON.stringify(greetingMessage))
+            this.hasRequestedInitialGreeting = true
             console.log('Sent initial greeting for scenario:', this.scenario)
           } else {
-            console.log('No scenario set, retrying in 200ms')
+            console.log('No scenario set, retrying in 300ms')
-            // Retry if scenario not set yet
+            // Retry if scenario not set yet with longer delay
-            setTimeout(() => this.requestInitialGreeting(), 200)
+            setTimeout(() => this.requestInitialGreeting(), 300)
           }
-        }, 100)
+        }, 200)  // Increased delay to ensure scenario is set
       }
     },
 
+    // Helper method to check if we should request greeting
+    checkAndRequestGreeting() {
+      if (this.websocket && this.websocket.readyState === WebSocket.OPEN && 
+          this.scenario && !this.hasRequestedInitialGreeting) {
+        this.requestInitialGreeting()
+      }
+    },
+
+    startConversation() {
+      // Hide the start button and show normal controls
+      this.showStartButton = false
+      
+      // Request the initial greeting now that user has interacted
+      this.checkAndRequestGreeting()
+    },
+
+    checkAndShowTutorial() {
+      // Check if user has seen tutorial before
+      const hasSeenTutorial = localStorage.getItem('streetLingo_tutorialSeen')
+      if (!hasSeenTutorial) {
+        this.showTutorial = true
+      }
+    },
+
+    closeTutorial() {
+      this.showTutorial = false
+      // Mark tutorial as seen
+      localStorage.setItem('streetLingo_tutorialSeen', 'true')
+    },
+
+
     setupAutoRecording(message) {
       // Wait for the audio element to be created in the DOM
       this.$nextTick(() => {
@@ -815,10 +905,6 @@ export default {
       }
     },
 
-    forceStopRecording() {
-      console.log('Force stopping recording')
-      this.stopRecording()
-    },
 
     startPauseDetection() {
       if (!this.suggestionsEnabled) {
@@ -940,6 +1026,9 @@ export default {
       }
       
       try {
+        // Set loading state
+        this.isLoadingFeedback = true
+        
         // Send conversation data to backend for feedback
         const response = await fetch(`${this.wsBaseUrl.replace('ws', 'http')}/api/conversation-feedback`, {
           method: 'POST',
@@ -982,6 +1071,9 @@ export default {
           suggestions: [],
           examples: []
         }
+      } finally {
+        // Clear loading state
+        this.isLoadingFeedback = false
       }
     },
 
@@ -1379,27 +1471,6 @@ export default {
   box-shadow: none;
 }
 
-.force-stop-btn {
-  background: #dc2626;
-  color: white;
-  border: none;
-  padding: 0.75rem;
-  border-radius: var(--radius-lg);
-  cursor: pointer;
-  font-size: 1rem;
-  transition: all 0.2s ease;
-  margin-left: 0.5rem;
-  width: 40px;
-  height: 40px;
-  display: flex;
-  align-items: center;
-  justify-content: center;
-}
-
-.force-stop-btn:hover {
-  background: #b91c1c;
-  transform: translateY(-1px);
-}
 
 .status {
   padding: 0.75rem;
@@ -1799,4 +1870,254 @@ export default {
   border-radius: var(--radius);
   font-weight: 500;
 }
+
+/* Feedback Loading Styles */
+.feedback-loading {
+  background: var(--surface);
+  border: 1px solid var(--border);
+  border-radius: var(--radius-lg);
+  padding: 1.5rem;
+  margin-bottom: 1.5rem;
+  box-shadow: var(--shadow-sm);
+  position: relative;
+  overflow: hidden;
+}
+
+.feedback-loading::before {
+  content: '';
+  position: absolute;
+  top: 0;
+  left: 0;
+  right: 0;
+  height: 3px;
+  background: var(--primary);
+}
+
+.loading-content {
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  gap: 1rem;
+  text-align: center;
+}
+
+.loading-content p {
+  margin: 0;
+  color: var(--text-light);
+  font-size: 1rem;
+  font-weight: 500;
+}
+
+.loading-spinner {
+  width: 24px;
+  height: 24px;
+  border: 3px solid var(--surface-alt);
+  border-top: 3px solid var(--primary);
+  border-radius: 50%;
+  animation: spin 1s linear infinite;
+}
+
+@keyframes spin {
+  0% { transform: rotate(0deg); }
+  100% { transform: rotate(360deg); }
+}
+
+/* Start Conversation Button Styles */
+.start-conversation-btn {
+  background: var(--primary);
+  color: white;
+  border: none;
+  padding: 1rem 2rem;
+  border-radius: var(--radius-lg);
+  cursor: pointer;
+  font-family: 'DM Sans', sans-serif;
+  font-size: 0.9rem;
+  font-weight: 500;
+  transition: all 0.3s cubic-bezier(0.4, 0, 0.2, 1);
+  box-shadow: var(--shadow);
+  display: flex;
+  align-items: center;
+  gap: 0.5rem;
+  min-width: 200px;
+  justify-content: center;
+}
+
+.start-conversation-btn:hover:not(:disabled) {
+  background: var(--primary-dark);
+  transform: translateY(-2px);
+  box-shadow: var(--shadow-lg);
+}
+
+.start-conversation-btn:disabled {
+  background: var(--text-muted);
+  cursor: not-allowed;
+  transform: none;
+  box-shadow: var(--shadow-sm);
+}
+
+/* Tutorial Popup Styles */
+.tutorial-overlay {
+  position: fixed;
+  top: 0;
+  left: 0;
+  right: 0;
+  bottom: 0;
+  background: rgba(0, 0, 0, 0.6);
+  display: flex;
+  align-items: center;
+  justify-content: center;
+  z-index: 9999;
+  backdrop-filter: blur(4px);
+}
+
+.tutorial-popup {
+  background: var(--surface);
+  border-radius: var(--radius-lg);
+  box-shadow: var(--shadow-lg);
+  max-width: 500px;
+  width: 90%;
+  max-height: 80vh;
+  overflow-y: auto;
+  animation: tutorialSlideIn 0.3s ease-out;
+}
+
+@keyframes tutorialSlideIn {
+  from {
+    opacity: 0;
+    transform: translateY(-20px) scale(0.95);
+  }
+  to {
+    opacity: 1;
+    transform: translateY(0) scale(1);
+  }
+}
+
+.tutorial-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: center;
+  padding: 1.5rem 1.5rem 1rem 1.5rem;
+  border-bottom: 1px solid var(--border);
+}
+
+.tutorial-header h3 {
+  margin: 0;
+  color: var(--text);
+  font-size: 1.3rem;
+  font-weight: 700;
+}
+
+.tutorial-close {
+  background: none;
+  border: none;
+  font-size: 1.2rem;
+  cursor: pointer;
+  color: var(--text-muted);
+  padding: 0.25rem;
+  border-radius: var(--radius);
+  transition: all 0.2s ease;
+}
+
+.tutorial-close:hover {
+  background: var(--surface-alt);
+  color: var(--text);
+}
+
+.tutorial-content {
+  padding: 1.5rem;
+}
+
+.tutorial-step {
+  display: flex;
+  align-items: flex-start;
+  gap: 1rem;
+  margin-bottom: 1.5rem;
+}
+
+.tutorial-step:last-child {
+  margin-bottom: 0;
+}
+
+.tutorial-icon {
+  font-size: 2rem;
+  flex-shrink: 0;
+}
+
+.tutorial-text {
+  flex: 1;
+}
+
+.tutorial-text strong {
+  display: block;
+  color: var(--text);
+  font-size: 1rem;
+  font-weight: 600;
+  margin-bottom: 0.5rem;
+}
+
+.tutorial-text p {
+  margin: 0;
+  color: var(--text-light);
+  font-size: 0.9rem;
+  line-height: 1.4;
+}
+
+.tutorial-footer {
+  padding: 1rem 1.5rem 1.5rem 1.5rem;
+  border-top: 1px solid var(--border);
+  display: flex;
+  justify-content: center;
+}
+
+.tutorial-got-it {
+  background: var(--primary);
+  color: white;
+  border: none;
+  padding: 0.75rem 2rem;
+  border-radius: var(--radius-lg);
+  cursor: pointer;
+  font-family: 'DM Sans', sans-serif;
+  font-size: 0.9rem;
+  font-weight: 600;
+  transition: all 0.2s ease;
+  box-shadow: var(--shadow-sm);
+}
+
+.tutorial-got-it:hover {
+  background: var(--primary-dark);
+  transform: translateY(-1px);
+  box-shadow: var(--shadow);
+}
+
+@media (max-width: 768px) {
+  .tutorial-popup {
+    width: 95%;
+    max-height: 85vh;
+  }
+  
+  .tutorial-header {
+    padding: 1rem 1rem 0.75rem 1rem;
+  }
+  
+  .tutorial-header h3 {
+    font-size: 1.1rem;
+  }
+  
+  .tutorial-content {
+    padding: 1rem;
+  }
+  
+  .tutorial-step {
+    gap: 0.75rem;
+    margin-bottom: 1rem;
+  }
+  
+  .tutorial-icon {
+    font-size: 1.5rem;
+  }
+  
+  .tutorial-footer {
+    padding: 0.75rem 1rem 1rem 1rem;
+  }
+}
 </style>

backend/languages/german/models.py

@@ -227,52 +227,114 @@ BIERGARTEN_PERSONALITIES = {
     )
 }
 
-UBAHN_PERSONALITIES = {
+BER_AIRPORT_PERSONALITIES = {
-    "bvg_info": GermanPersonality(
+    "airport_info": GermanPersonality(
         character_type=CharacterType.SERVICE_WORKER,
-        name="BVG Mitarbeiter",
+        name="Flughafen-Mitarbeiter",
-        gender=Gender.MALE,
+        gender=Gender.FEMALE,
         tone=PersonalityTone.HELPFUL,
         age_range="young",
-        background="Helpful BVG information staff at U-Bahn station",
+        background="Helpful airport staff at BER who knows all train connections",
         typical_phrases=[
-            "Kann ich Ihnen helfen?",
+            "Guten Tag! Kann ich Ihnen helfen?",
-            "Wohin möchten Sie denn?",
+            "Ach, nach Potsdamer Platz?",
-            "Nehmen Sie die U6 Richtung...",
+            "Ja, die Züge haben heute Verspätung",
-            "Steigen Sie an... um",
+            "Schauen Sie mal auf die Anzeigetafel",
-            "Das sind drei Stationen",
+            "Gleis 3 oder 4",
-            "Brauchen Sie eine Fahrkarte?",
+            "Der Airport Express fährt alle 20 Minuten",
-            "Zone AB reicht",
+            "Das dauert etwa 40 Minuten",
-            "Gute Fahrt!"
+            "Nehmen Sie den FEX oder die RE7",
+            "Steigen Sie am Hauptbahnhof um"
         ],
-        response_style="Professional and helpful with public transport",
+        response_style="Professional and understanding about travel delays",
-        location_context="U-Bahn station information desk",
+        location_context="BER Airport terminal information desk",
-        scenario_title="U-Bahn Help",
+        scenario_title="BER Airport Train Help",
-        scenario_description="You're at a Berlin U-Bahn station asking for directions and transport information. Practice asking about public transport in German.",
+        scenario_description="You've just arrived at BER airport and need to get to your hotel at Potsdamer Platz, but Google Maps shows train delays. Ask for help with train connections and platforms.",
-        scenario_challenge="Understanding German public transport terminology, directions, and ticket system.",
+        scenario_challenge="Understanding German train terminology, dealing with delays, and navigating airport-to-city connections.",
-        scenario_goal="Get directions and buy appropriate ticket",
+        scenario_goal="Get clear directions to Potsdamer Platz despite train delays",
         goal_items=[
             GoalItem(
-                id="ask_directions",
+                id="explain_destination",
-                description="Ask for directions (Nach dem Weg fragen)"
+                description="Explain you need to get to Potsdamer Platz (Ziel erklären)"
             ),
             GoalItem(
-                id="buy_ticket",
+                id="ask_about_delays",
-                description="Buy appropriate ticket (Passende Fahrkarte kaufen)"
+                description="Ask about train delays (Nach Verspätungen fragen)"
+            ),
+            GoalItem(
+                id="get_platform_info",
+                description="Find out which platform and train to take (Gleis und Zug erfragen)"
             )
         ],
         helpful_phrases=[
-            HelpfulPhrase(native="Wie komme ich nach...?", english="How do I get to...?"),
+            HelpfulPhrase(native="Ich muss nach Potsdamer Platz", english="I need to get to Potsdamer Platz"),
-            HelpfulPhrase(native="Welche Linie muss ich nehmen?", english="Which line do I need to take?"),
+            HelpfulPhrase(native="Welcher Zug fährt dorthin?", english="Which train goes there?"),
-            HelpfulPhrase(native="Wo muss ich umsteigen?", english="Where do I need to change?"),
+            HelpfulPhrase(native="Haben die Züge Verspätung?", english="Are the trains delayed?"),
-            HelpfulPhrase(native="Wie viele Stationen?", english="How many stations?"),
+            HelpfulPhrase(native="Von welchem Gleis fährt der Zug?", english="From which platform does the train leave?"),
-            HelpfulPhrase(native="Welche Fahrkarte brauche ich?", english="Which ticket do I need?"),
+            HelpfulPhrase(native="Wie lange dauert die Fahrt?", english="How long does the journey take?"),
-            HelpfulPhrase(native="Einzelfahrkarte", english="Single ticket"),
+            HelpfulPhrase(native="Muss ich umsteigen?", english="Do I need to change trains?"),
-            HelpfulPhrase(native="Tageskarte", english="Day ticket"),
+            HelpfulPhrase(native="Airport Express", english="Airport Express"),
-            HelpfulPhrase(native="Richtung", english="Direction")
+            HelpfulPhrase(native="Anzeigetafel", english="Display board"),
+            HelpfulPhrase(native="Hauptbahnhof", english="Main station")
         ],
         is_helpful=True,
-        is_talkative=False,
+        is_talkative=True,
+        uses_slang=False
+    )
+}
+
+ARZT_PERSONALITIES = {
+    "dr_muller": GermanPersonality(
+        character_type=CharacterType.OFFICIAL,
+        name="Dr. Müller",
+        gender=Gender.MALE,
+        tone=PersonalityTone.FORMAL,
+        age_range="middle-aged",
+        background="General practitioner in Berlin who is thorough and caring",
+        typical_phrases=[
+            "Guten Tag, setzen Sie sich bitte.",
+            "Was führt Sie zu mir?",
+            "Seit wann haben Sie diese Beschwerden?",
+            "Wo genau tut es weh?",
+            "Auf einer Skala von 1 bis 10?",
+            "Ich schaue mir das mal an",
+            "Ich gebe Ihnen eine Überweisung",
+            "Nehmen Sie diese Medikamente",
+            "Gute Besserung!"
+        ],
+        response_style="Professional, caring, asks detailed questions about symptoms",
+        location_context="Doctor's office in Berlin",
+        scenario_title="At the Doctor",
+        scenario_description="You've been having leg pain for the past week and need to see a doctor. Practice describing symptoms and asking for a specialist referral in German.",
+        scenario_challenge="Understanding medical terminology, describing pain and symptoms accurately, and navigating the German healthcare system.",
+        scenario_goal="Describe leg pain and get specialist referral",
+        goal_items=[
+            GoalItem(
+                id="describe_symptoms",
+                description="Describe your leg pain symptoms (Beinschmerzen beschreiben)"
+            ),
+            GoalItem(
+                id="explain_duration",
+                description="Explain how long you've had the pain (Dauer der Schmerzen erklären)"
+            ),
+            GoalItem(
+                id="request_referral",
+                description="Ask for a specialist referral (Überweisung zum Facharzt bitten)"
+            )
+        ],
+        helpful_phrases=[
+            HelpfulPhrase(native="Mir tut das Bein weh", english="My leg hurts"),
+            HelpfulPhrase(native="Seit einer Woche", english="For a week"),
+            HelpfulPhrase(native="Die Schmerzen sind hier", english="The pain is here"),
+            HelpfulPhrase(native="Es tut sehr weh", english="It hurts a lot"),
+            HelpfulPhrase(native="Können Sie mir eine Überweisung geben?", english="Can you give me a referral?"),
+            HelpfulPhrase(native="Ich brauche einen Facharzt", english="I need a specialist"),
+            HelpfulPhrase(native="Orthopäde", english="Orthopedist"),
+            HelpfulPhrase(native="Schmerztabletten", english="Painkillers"),
+            HelpfulPhrase(native="Krankenschein", english="Sick note")
+        ],
+        is_helpful=True,
+        is_talkative=True,
         uses_slang=False
     )
 }
@@ -283,5 +345,6 @@ SCENARIO_PERSONALITIES = {
     "wg_viewing": WG_PERSONALITIES,
     "burgeramt": BURGERAMT_PERSONALITIES,
     "biergarten": BIERGARTEN_PERSONALITIES,
-    "ubahn": UBAHN_PERSONALITIES
+    "ber_airport": BER_AIRPORT_PERSONALITIES,
+    "arzt": ARZT_PERSONALITIES
 }

backend/languages/german/services.py

@@ -41,10 +41,16 @@ class GermanTextToSpeechService(TextToSpeechService):
                 "pitch": None,
                 "ssml_gender": texttospeech.SsmlVoiceGender.MALE,
             },
-            "BVG Mitarbeiter": {
+            "Flughafen-Mitarbeiter": {
-                "name": "de-DE-Chirp3-HD-Fenrir",  # Professional male voice
+                "name": "de-DE-Chirp3-HD-Leda",  # Professional female voice
                 "speaking_rate": 0.95,
                 "pitch": None,
+                "ssml_gender": texttospeech.SsmlVoiceGender.FEMALE,
+            },
+            "Dr. Müller": {
+                "name": "de-DE-Chirp3-HD-Fenrir",  # Professional male voice for doctor
+                "speaking_rate": 0.9,
+                "pitch": None,
                 "ssml_gender": texttospeech.SsmlVoiceGender.MALE,
             }
         }
@@ -129,8 +135,10 @@ class GermanConversationFlowService(BaseConversationFlowService):
             detected_scenario = "burgeramt"
         elif "biergarten" in context_lower or "beer" in context_lower or "restaurant" in context_lower:
             detected_scenario = "biergarten"
-        elif "ubahn" in context_lower or "u-bahn" in context_lower or "transport" in context_lower:
+        elif "ber" in context_lower or "airport" in context_lower or "flughafen" in context_lower or "potsdamer" in context_lower:
-            detected_scenario = "ubahn"
+            detected_scenario = "ber_airport"
+        elif "arzt" in context_lower or "doctor" in context_lower or "medical" in context_lower:
+            detected_scenario = "arzt"
         else:
             detected_scenario = "spati"  # Default to späti
             

backend/main.py

@@ -373,10 +373,11 @@ Focus on common German language learning areas:
             target_language = "Indonesian"
             language_specific_feedback = """
 Focus on common Indonesian language learning areas:
-- Formal vs informal language (using proper pronouns)
+- Using everyday, natural Indonesian words and expressions
+- Sounding more natural and conversational (not textbook formal)
+- Common Indonesian idioms and colloquial expressions
 - Sentence structure and word order
-- Common Indonesian expressions
+- Building confidence in casual conversation
-- Politeness levels and cultural context
 """
         
         feedback_prompt = f"""You are an encouraging {target_language} language teacher. A student has just finished a conversation practice session in a {request.scenario} scenario. 
@@ -386,35 +387,77 @@ Here's their conversation:
 
 {language_specific_feedback}
 
+MANDATORY ANALYSIS: Before providing feedback, carefully examine each thing the student said for language issues. Look for:
+1. Unnatural phrasing or word choices
+2. Grammar mistakes or awkward constructions  
+3. Word order problems
+4. Missing words that would make meaning clearer
+5. Overly formal or informal expressions for the context
+
+If you find ANY of these issues in their actual speech, you MUST provide specific suggestions and examples. Do not give empty suggestions/examples arrays unless their language was genuinely perfect.
+
 Provide helpful, encouraging feedback as a JSON object with:
 - "encouragement": A positive, motivating message about their effort (2-3 sentences)
-- "suggestions": Array of 2-3 objects with:
+- "suggestions": Array of 0-3 objects with:
   - "category": Area of improvement (e.g., "Pronunciation", "Grammar", "Vocabulary")
-  - "tip": Specific, actionable advice
+  - "tip": Specific, actionable advice based ONLY on what they actually said in the conversation
-- "examples": Array of 1-2 objects with:
+- "examples": Array of 0-2 objects with:
   - "original": Something they actually said (from the conversation)
   - "improved": A better way to say it
   - "reason": Brief explanation of why it's better
 
+CRITICAL REQUIREMENT: You MUST analyze the student's actual words and phrases for improvement opportunities. Common Indonesian learner issues to look for:
+
+1. **Word Order**: "ayam indomi" should be "indomie ayam" (flavor comes after product)
+2. **Phrasing**: "beli ayam indomi satu sama mau minum" is awkward - should be "mau beli indomie ayam sama minum"  
+3. **Missing Words**: "mau stroberi ultra milk" missing "yang" (mau yang stroberi)
+4. **Unclear Intent**: "saya beli minum" should be "saya mau beli minum" (clearer intention)
+
+Do NOT give empty suggestions/examples unless the conversation was genuinely flawless. If there are language issues (which there usually are), provide specific, helpful corrections.
+
 Make it encouraging and supportive, focusing on growth rather than criticism. If they did well, focus on areas to sound more natural or confident.
 
-Example format:
+For Indonesian specifically:
+- Focus on everyday conversational language rather than formal politeness
+- Emphasize natural, casual expressions that locals actually use
+- Include specific word examples in your tips
+- Avoid focusing on formal grammar rules - prioritize natural communication
+
+Example format for good conversation (no meaningful improvements needed):
+{{
+  "encouragement": "Fantastic job in your conversation practice! You really engaged well and made your choices clear. Keep up the great work, and your confidence will only grow!",
+  "suggestions": [],
+  "examples": []
+}}
+
+Example format for conversation with meaningful improvements:
 {{
   "encouragement": "You did a great job engaging in this conversation! Your effort to communicate is really paying off.",
   "suggestions": [
     {{
-      "category": "Vocabulary",
+      "category": "Word Order",
-      "tip": "Try using more common everyday words to sound more natural"
+      "tip": "In Indonesian, the product name comes first, then the flavor - so 'indomie ayam' instead of 'ayam indomi'"
+    }},
+    {{
+      "category": "Phrasing",
+      "tip": "When expressing what you want to buy, use 'mau beli' to be clearer about your intention"
     }}
   ],
   "examples": [
     {{
-      "original": "I want to purchase this item",
+      "original": "beli ayam indomi satu sama mau minum",
-      "improved": "I'd like to buy this",
+      "improved": "mau beli indomie ayam satu sama minum",
-      "reason": "Sounds more natural and conversational"
+      "reason": "Better word order and clearer intention - 'mau beli' shows you want to buy"
+    }},
+    {{
+      "original": "mau stroberi ultra milk",
+      "improved": "mau ultra milk yang stroberi",
+      "reason": "Adding 'yang' makes it clearer which flavor you want"
     }}
   ]
-}}"""
+}}
+
+IMPORTANT: Analyze the student's actual phrases from the conversation above and provide specific corrections for any unnatural or incorrect expressions. Don't give empty arrays unless their Indonesian was perfect."""
         
         response = client.chat.completions.create(
             model=config.OPENAI_MODEL,
@@ -500,6 +543,7 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
     transcript_repeat_count = 0
     last_transcript = ""
     high_confidence_count = 0
+    session_processed = False  # Flag to prevent duplicate processing
     
     import uuid
     session_id = str(uuid.uuid4())
@@ -518,6 +562,7 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                 chunk_count = 0
                 latest_transcript = ""
                 recording_start_time = time.time()
+                session_processed = False  # Reset processing flag for new session
                 logger.info("Started recording session")
                 
             elif message["type"] == "conversation_reset":
@@ -527,17 +572,39 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
             elif message["type"] == "audio_chunk":
                 if is_recording:
                     # Check for recording timeout
-                    if recording_start_time and time.time() - recording_start_time > max_recording_duration:
+                    if recording_start_time and time.time() - recording_start_time > max_recording_duration and not session_processed:
                         logger.warning("Recording timeout reached, auto-stopping")
+                        is_recording = False
+                        session_processed = True  # Mark as processed to prevent duplicates
+                        
                         # Send timeout notification to frontend
                         timeout_notification = {
                             "type": "recording_timeout",
                             "message": "Recording stopped due to timeout"
                         }
                         await websocket.send_text(json.dumps(timeout_notification))
-                        # Force audio_end processing
+                        
-                        message = {"type": "audio_end", "scenario_context": message.get("scenario_context", "")}
+                        # Process final transcript if available
-                        # Don't return, let it fall through to audio_end processing
+                        if latest_transcript.strip():
+                            transcription_result = {
+                                "type": "transcription",
+                                "transcript": latest_transcript,
+                                "is_final": True,
+                                "confidence": 0.8
+                            }
+                            await websocket.send_text(json.dumps(transcription_result))
+                            
+                            # Process AI response
+                            logger.info("Getting AI response after timeout...")
+                            ai_response = await session_conversation_service.process_conversation_flow_fast(
+                                latest_transcript, 
+                                message.get("scenario_context", "")
+                            )
+                            logger.info(f"AI response: {ai_response.get('text', 'No text')}")
+                            await websocket.send_text(json.dumps(ai_response))
+                        
+                        audio_buffer.clear()
+                        continue  # Skip further processing for this message
                     else:
                         audio_data = base64.b64decode(message["audio"])
                         logger.info(f"Received audio chunk: {len(audio_data)} bytes")
@@ -548,8 +615,8 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                         # Process chunk for real-time transcription
                         chunk_count += 1
                         try:
-                            # Only process every 8th chunk to reduce log spam and API calls
+                            # Only process every 6th chunk for faster response (reduced from 8th)
-                            if chunk_count % 8 == 0 and len(audio_buffer) >= 19200:  # ~0.4 seconds of audio at 48kHz
+                            if chunk_count % 6 == 0 and len(audio_buffer) >= 14400:  # ~0.3 seconds of audio at 48kHz
                                 recognition_audio = speech.RecognitionAudio(content=bytes(audio_buffer))
                                 response = session_conversation_service.stt_service.client.recognize(
                                     config=session_conversation_service.stt_service.recognition_config,
@@ -561,7 +628,7 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                                     confidence = response.results[0].alternatives[0].confidence
                                     
                                     # Store transcript if confidence is reasonable (lowered for speed)
-                                    if confidence > 0.6:
+                                    if confidence > 0.4:  # Lowered threshold for faster processing
                                         latest_transcript = transcript  # Store latest transcript
                                         
                                         # Check for repeated high-confidence transcripts
@@ -570,13 +637,13 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                                                 high_confidence_count += 1
                                                 logger.info(f"Repeated high confidence transcript #{high_confidence_count}: '{transcript}' (confidence: {confidence})")
                                                 
-                                                # If we've seen the same high-confidence transcript 4+ times, auto-stop
+                                                # If we've seen the same high-confidence transcript 3+ times, auto-stop (reduced from 4)
-                                                if high_confidence_count >= 4:
+                                                if high_confidence_count >= 3 and not session_processed:
                                                     logger.info("Auto-stopping recording due to repeated high-confidence transcript")
                                                     is_recording = False
+                                                    session_processed = True  # Mark as processed to prevent duplicates
+                                                    
                                                     # Send final processing message
-                                                    final_message = {"type": "audio_end", "scenario_context": message.get("scenario_context", "")}
-                                                    # Process immediately without waiting for more chunks
                                                     await websocket.send_text(json.dumps({
                                                         "type": "transcription",
                                                         "transcript": transcript,
@@ -640,15 +707,22 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
             
             elif message["type"] == "audio_end":
                 is_recording = False
+                
+                # Check if this session was already processed (e.g., by auto-stop logic)
+                if session_processed:
+                    logger.info("Audio session already processed, skipping duplicate processing")
+                    continue
+                    
                 final_transcript = ""
                 
                 # Use latest interim transcript if available for faster response
                 logger.info(f"Checking latest_transcript: '{latest_transcript}'")
-                if latest_transcript.strip():
+                if latest_transcript.strip() and len(latest_transcript.strip()) > 3:  # More aggressive check
                     final_transcript = latest_transcript
                     logger.info(f"Using latest interim transcript: '{final_transcript}'")
+                    session_processed = True  # Mark as processed to prevent duplicates
                     
-                    # Send final transcription immediately
+                    # Send final transcription immediately - no "Processing..." delay
                     transcription_result = {
                         "type": "transcription",
                         "transcript": final_transcript,
@@ -657,8 +731,8 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                     }
                     await websocket.send_text(json.dumps(transcription_result))
                     
-                    # Process AI response with faster flow
+                    # Process AI response with faster flow - start immediately
-                    logger.info("Getting AI response...")
+                    logger.info("Getting AI response immediately...")
                     ai_response = await session_conversation_service.process_conversation_flow_fast(
                         final_transcript, 
                         message.get("scenario_context", "")
@@ -673,6 +747,7 @@ async def websocket_speech_endpoint(websocket: WebSocket, language: str):
                 elif len(audio_buffer) > 0:
                     # Fallback to full transcription if no interim results
                     logger.info(f"Processing final audio buffer: {len(audio_buffer)} bytes")
+                    session_processed = True  # Mark as processed to prevent duplicates
                     try:
                         recognition_audio = speech.RecognitionAudio(content=bytes(audio_buffer))
                         response = session_conversation_service.stt_service.client.recognize(