addedtts

app.py

@@ -1,23 +1,33 @@
 import gradio as gr
 from asr import transcribe_audio
 from llm_agent import get_llm_reply
+from tts import generate_speech
 
-def voice_to_reply(audio):
+def voice_agent(audio):
     if audio is None:
-        return "No audio received!", "..."
-    text = transcribe_audio(audio)
-    reply = get_llm_reply(text)
-    return text, reply
+        return "No audio received!", "...", None
+    
+    # ASR
+    user_text = transcribe_audio(audio)
+    
+    # LLM
+    bot_reply = get_llm_reply(user_text)
+    
+    # TTS
+    audio_reply = generate_speech(bot_reply)
+    
+    return user_text, bot_reply, audio_reply
 
 ui = gr.Interface(
-    fn=voice_to_reply,
-    inputs=gr.Audio(type="filepath", label="Upload voice (.wav or .mp3)"),
+    fn=voice_agent,
+    inputs=gr.Audio(type="filepath", label="🎙️ Upload Voice (.wav or .mp3)"),
     outputs=[
         gr.Textbox(label="📝 Transcribed Text"),
-        gr.Textbox(label="🤖 LLM Reply")
+        gr.Textbox(label="🤖 LLM Reply"),
+        gr.Audio(type="filepath", label="🔊 AI Voice Response")
     ],
-    title="VoiceFreight AI v1: ASR + LLM",
-    description="Upload a voice file. ASR + TinyLlama gives you a response!"
+    title="VoiceFreight AI v1.0",
+    description="Upload a voice file. Get a spoken reply from an AI freight agent!"
 )
 
 if __name__ == "__main__":

tts.py

@@ -0,0 +1,8 @@
+from TTS.api import TTS
+
+
+tts = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False, gpu=False)
+
+def generate_speech(text, filename="output.wav"):
+    tts.tts_to_file(text=text, file_path=filename)
+    return filename