Update app.py

app.py

@@ -23,16 +23,15 @@ tts_model = get_tts_model()
 conversations: dict[str, list[dict[str, str]]] = {}
 
 
+llm_client = InferenceClient(provider="groq")
+
+
 def response(
     audio: tuple[int, np.ndarray],
     hf_token: str | None,
 ):
     if hf_token is None:
         raise WebRTCError("HF Token is required")
-    llm_client = InferenceClient(
-        provider="groq",
-        api_key=hf_token,
-    )
     context = get_current_context()
     if context.webrtc_id not in conversations:
         conversations[context.webrtc_id] = [
@@ -55,6 +54,7 @@ def response(
         messages=messages,  # type: ignore
         max_tokens=1024,
         stream=True,
+        token=hf_token,
     )
 
     output_text = ""