Add explicit user gesture to play intro TTS in demo interface, enhancing cross-browser compatibility. Refactor initialization to separate TTS loading logic.

app.py

@@ -587,6 +587,15 @@ def _on_load_init_with_tts(tts_on: bool):
     tts_path = synthesize_tts(chat_history[-1][1]) if bool(tts_on) else None
     return chat_history, scores_state, meta_state, finished_state, turns_state, tts_path
 
+
+def _play_intro_tts(tts_on: bool):
+    if not bool(tts_on):
+        return None
+    try:
+        return synthesize_tts(INTRO_MESSAGE)
+    except Exception:
+        return None
+
 def create_demo():
     with gr.Blocks(
         theme=gr.themes.Soft(),
@@ -632,6 +641,7 @@ def create_demo():
             Tap on 'Record' to start speaking, then tap on 'Stop' to stop recording.
             """
         )
+        intro_play_btn = gr.Button("▶️ Play Intro", variant="secondary")
 
         with gr.Tabs():
             with gr.TabItem("Main"):
@@ -664,8 +674,11 @@ def create_demo():
         finished_state = gr.State()
         turns_state = gr.State()
 
-        # Initialize on load and optionally speak the intro message
-        demo.load(_on_load_init_with_tts, inputs=[tts_enable], outputs=[chatbot, scores_state, meta_state, finished_state, turns_state, tts_audio_main])
+        # Initialize on load (no autoplay due to browser policies)
+        demo.load(_on_load_init, inputs=None, outputs=[chatbot, scores_state, meta_state, finished_state, turns_state])
+
+        # Explicit user gesture to play intro TTS (works across browsers)
+        intro_play_btn.click(fn=_play_intro_tts, inputs=[tts_enable], outputs=[tts_audio_main])
 
         # Wire interactions
         audio_main.stop_recording(