Update app.py

app.py

@@ -8,10 +8,8 @@ def _load_key() -> str:
     key = os.environ.get("GROQ_API_KEY") or os.environ.get("groq_api_key")
     if not key:
         raise RuntimeError(
-            "Groq API key not found. In your Space, add a secret named 'groq_api_key' "
-            "or set env var GROQ_API_KEY."
+            "Groq API key not found. In your Space settings -> Secrets, add 'groq_api_key'."
         )
-    # Ensure SDK can also read it from the standard name
     os.environ["GROQ_API_KEY"] = key
     return key
 
@@ -26,8 +24,7 @@ def transcribe_audio(audio_path: str, model: str = "whisper-large-v3") -> str:
             model=model,
             response_format="verbose_json",
         )
-    text = getattr(resp, "text", "") or ""
-    return text.strip()
+    return (getattr(resp, "text", "") or "").strip()
 
 def stream_answer(prompt_text: str,
                   model: str = "llama-3.1-8b-instant",
@@ -67,7 +64,7 @@ def text_to_speech(text: str,
         input=tts_input,
     )
     out_path = os.path.join(tempfile.gettempdir(), f"answer_{int(time.time())}.{fmt}")
-    # Groq SDK returns BinaryAPIResponse. Use write_to_file, not stream_to_file.
+    # BinaryAPIResponse uses write_to_file in Groq SDK
     resp.write_to_file(out_path)
     return out_path
 
@@ -109,7 +106,7 @@ def run_pipeline(audio_file, typed_question, llm_model, voice_name):
         short_tb = "\n".join(traceback.format_exc().splitlines()[-6:])
         help_tip = (
             "\nTips:\n"
-            "- Check your Groq key in Space secrets.\n"
+            "- Check Space secret 'groq_api_key'.\n"
             "- Try a shorter audio clip.\n"
             "- Verify model names.\n"
             "- Confirm requirements installed."
@@ -158,5 +155,7 @@ with gr.Blocks(title="Audio Q&A with Groq") as demo:
         return "", "", None, ""
     clear_btn.click(fn=clear_all, inputs=None, outputs=[transcript_box, answer_box, answer_audio, status_md])
 
-# Use queue for better concurrency on Spaces
-demo.queue(concurrency_count=2).launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860)))
+if __name__ == "__main__":
+    # On HF Spaces you can simply do demo.launch()
+    # Queue enables generator streaming without extra args in Gradio v4
+    demo.queue().launch()