fix: create playback AudioContext in user gesture for mobile support

- Playback AudioContext created during startConversation (button click)
- Removed sampleRate constraint from getUserMedia (let browser choose)
- Added audio chunk logging for debugging

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

src/components/configurator/VoiceAgentProvider.tsx

@@ -206,10 +206,7 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
   );
 
   const playAudioChunk = useCallback((base64Audio: string) => {
-    if (!playbackContextRef.current) {
-      playbackContextRef.current = new AudioContext({ sampleRate: 24000 });
-      nextStartTimeRef.current = playbackContextRef.current.currentTime;
-    }
+    if (!playbackContextRef.current) return;
     const ctx = playbackContextRef.current;
     const int16 = base64ToInt16(base64Audio);
     const float32 = int16ToFloat32(int16);
@@ -248,13 +245,17 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
       const { apiKey, model, config } = tokenData;
 
       const stream = await navigator.mediaDevices.getUserMedia({
-        audio: { sampleRate: 16000, channelCount: 1, echoCancellation: true, noiseSuppression: true },
+        audio: { channelCount: 1, echoCancellation: true, noiseSuppression: true },
       });
       mediaStreamRef.current = stream;
 
-      // Create AudioContext for mic capture (must be in user gesture handler)
+      // Create AudioContexts during user gesture (required on mobile)
       const audioCtx = new AudioContext({ sampleRate: 16000 });
       audioContextRef.current = audioCtx;
+
+      // Playback context MUST be created here (user gesture) for mobile
+      playbackContextRef.current = new AudioContext({ sampleRate: 24000 });
+      nextStartTimeRef.current = playbackContextRef.current.currentTime;
       const source = audioCtx.createMediaStreamSource(stream);
 
       const analyser = audioCtx.createAnalyser();
@@ -351,10 +352,12 @@ export default function VoiceAgentProvider({ locale, children }: VoiceAgentProvi
           const parts = msg.serverContent.modelTurn?.parts;
           if (parts) {
             for (const part of parts) {
-              if (part.inlineData?.mimeType?.startsWith('audio/')) {
+              if (part.inlineData) {
+                console.log('[VoiceAgent] Audio chunk received, mime:', part.inlineData.mimeType, 'len:', part.inlineData.data?.length);
                 playAudioChunk(part.inlineData.data);
               }
               if (part.text) {
+                console.log('[VoiceAgent] Text:', part.text);
                 addTranscript('agent', part.text);
               }
             }