Create app.py

app.py

@@ -0,0 +1,40 @@
+import gradio as gr
+from TTS.api import TTS
+
+# Load YourTTS / XTTS v2 model from Hugging Face
+model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
+tts = TTS(model_name)
+
+def generate_speech(text, emotion, ref_audio=None):
+    output_path = "output.wav"
+    
+    # If a reference voice is provided, use it for cloning
+    if ref_audio:
+        tts.tts_to_file(
+            text=text,
+            file_path=output_path,
+            speaker_wav=ref_audio,
+            emotion=emotion
+        )
+    else:
+        tts.tts_to_file(
+            text=text,
+            file_path=output_path,
+            emotion=emotion
+        )
+    return output_path
+
+
+iface = gr.Interface(
+    fn=generate_speech,
+    inputs=[
+        gr.Textbox(label="Text to Speak"),
+        gr.Dropdown(choices=["neutral", "happy", "sad", "angry"], label="Emotion"),
+        gr.Audio(source="upload", type="filepath", label="Reference Voice (optional)")
+    ],
+    outputs=gr.Audio(type="filepath"),
+    title="Voice Cloning with Emotions",
+    description="Upload a short voice sample, type text, pick emotion → Get speech in that cloned voice."
+)
+
+iface.launch()