Update tts_service.py

services/tts_service.py

@@ -1,9 +1,10 @@
 from services.gemini_client import get_gemini_client
 from google.genai import types
+import base64
 
 async def generate_tts(text: str) -> bytes:
     """
-    Convert text to speech using Gemini API
+    Convert text to speech using Gemini API and return valid WAV bytes
     """
     client = get_gemini_client()
 
@@ -20,5 +21,8 @@ async def generate_tts(text: str) -> bytes:
         ),
     )
 
-    # Return raw audio bytes
-    return response.candidates[0].content.parts[0].inline_data.data
+    # Gemini returns base64 audio, decode it to bytes
+    audio_base64 = response.candidates[0].content.parts[0].inline_data.data
+    audio_bytes = base64.b64decode(audio_base64)
+
+    return audio_bytes