Update app.py

app.py

@@ -134,27 +134,30 @@ def generate_response(transcribed_text):
     )
     return response.choices[0].message.content
 
-
 def inference(text):
    #client = openai.OpenAI(api_key='your_api_key')
    response = client.audio.speech.create(
-     model="tts-1",
-     voice="alloy", 
-     input=text
+       model="tts-1",
+       voice="alloy", 
+       input=text
    )
    output_file = "tts_output.mp3"
    response.stream_to_file(output_file)
    return output_file
+    
+def process_audio_and_respond(audio):
+   text = transcribe(audio)
+   response_text = generate_response(text)
+   output_file = inference(response_text)
+   return response_text, output_file
 
 demo = gr.Interface(
-    process_audio_and_respond,
-    gr.inputs.Audio(source="microphone", type="filepath", label="Bonyeza kitufe cha kurekodi na uliza swali lako"),
-    [gr.outputs.Textbox(label="Jibu (kwa njia ya maandishi)"), gr.outputs.Audio(type="filepath", label="Jibu kwa njia ya sauti (Bofya kusikiliza Jibu)")],
-    description="Uliza Swali kuhusu haki and sheria za udongo",
-    theme="compact",
-    layout="vertical",
-    allow_flagging=False,
-    live=True,
-)
-
-demo.launch()
+   process_audio_and_respond,
+   gr.inputs.Audio(source="microphone", type="filepath", label="Bonyeza kitufe cha kurekodi na uliza swali lako"),
+   [gr.outputs.Textbox(label="Jibu (kwa njia ya maandishi)"), gr.outputs.Audio(type="filepath", label="Jibu kwa njia ya sauti (Bofya kusikiliza Jibu)")],
+   description="Uliza Swali kuhusu haki and sheria za udongo",
+   theme="compact",
+   layout="vertical",
+   allow_flagging=False,
+   live=True,
+)