Update app.py

app.py

@@ -10,6 +10,7 @@ from Gradio_UI import GradioUI
 from kokoro import KPipeline
 import soundfile as sf
 import os
+import numpy as np
 
 # Below is an example of a tool that does nothing. Amaze us with your creativity !
 @tool
@@ -36,20 +37,24 @@ def text_to_speech_kokoro(text: str, voice: str = 'af_heart', speed: float = 1.0
         speed: The speed of the speech (default is 1.0).
     
     Returns:
-        The filename of the generated audio file.
+        A tuple containing the sample rate and the generated audio data as a NumPy array.
     """
     try:
         # Generate speech audio
         generator = pipeline(text, voice=voice, speed=speed, split_pattern=r'\n+')
-        audio_files = []
+        audio_segments = []
 
-        # Save each audio segment to a file
-        for i, (gs, ps, audio) in enumerate(generator):
-            filename = f'output_{i}.wav'
-            sf.write(filename, audio, 24000)
-            audio_files.append(filename)
+         # Collect each audio segment
+        for _, _, audio in generator:
+            audio_segments.append(audio)
 
-        return f"Generated {len(audio_files)} audio file(s): {', '.join(audio_files)}"
+        # Concatenate all audio segments into a single array
+        if audio_segments:
+            full_audio = np.concatenate(audio_segments)
+            sample_rate = 24000  # Kokoro-82M outputs audio at 24 kHz
+            return sample_rate, full_audio
+        else:
+            return "No audio generated."
     except Exception as e:
         return f"Error generating speech: {str(e)}"