Update app.py

app.py

@@ -6,62 +6,56 @@ import whisper
 import torchaudio
 from groq import Groq
 
-
-# Load Whisper model
+# Load Whisper model for transcription
 whisper_model = whisper.load_model("tiny")
 
-# Groq API client
+# Initialize Groq LLM client
 groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
-st.set_page_config(page_title="Voice Cloner Chat", layout="centered")
-st.title("🎙️ Voice Cloner Chat App")
+st.set_page_config(page_title="Voice Chat", layout="centered")
+st.title("🎤 Voice Chat using Your Cloned Voice")
 
-# Session State
+# Store cloned voice path
 if "clone_path" not in st.session_state:
     st.session_state.clone_path = None
 
-st.sidebar.header("Voice Input")
-option = st.sidebar.radio("Choose input method", ["Upload Voice", "Record Voice"])
+st.sidebar.header("Setup Your Clone Voice")
+voice_option = st.sidebar.radio("Choose how to provide clone voice", ["Upload Voice", "Record Voice"])
 
-if option == "Upload Voice":
-    uploaded = st.sidebar.file_uploader("Upload voice sample", type=["wav", "mp3"])
+if voice_option == "Upload Voice":
+    uploaded = st.sidebar.file_uploader("Upload a voice sample", type=["wav", "mp3"])
     if uploaded:
-        voice_path = save_uploaded_audio(uploaded, "reference_voice.wav")
-        st.session_state.clone_path = voice_path
-        st.success("Voice uploaded and saved as clone voice.")
-
-if option == "Record Voice":
-    duration = st.sidebar.slider("Duration (seconds)", 3, 10, 5)
-    if st.sidebar.button("Record"):
-        from utils import record_audio
-        path = record_audio(duration=duration)
+        path = save_uploaded_audio(uploaded, "reference_voice.wav")
         st.session_state.clone_path = path
-        st.success("Voice recorded and saved as clone voice.")
+        st.success("Voice uploaded and saved as your clone voice.")
 
-st.divider()
-st.subheader("💬 Talk with AI using cloned voice")
+# --- Main conversation section ---
+st.subheader("🗣️ Ask something using your voice")
 
 user_voice = st.file_uploader("Upload your voice question", type=["wav", "mp3"])
 if user_voice:
     user_voice_path = save_uploaded_audio(user_voice, "user_question.wav")
     st.audio(user_voice_path)
 
-    # Transcribe question
+    # Step 1: Transcribe
+    st.info("Transcribing...")
     result = whisper_model.transcribe(user_voice_path)
     user_text = result["text"]
-    st.write(f"📝 Transcribed Text: {user_text}")
+    st.success(f"📝 You said: {user_text}")
 
-    # Generate response using Groq
+    # Step 2: Generate LLM response
+    st.info("Thinking...")
     response = groq_client.chat.completions.create(
         model="llama3-8b-8192",
         messages=[{"role": "user", "content": user_text}]
     )
     reply = response.choices[0].message.content
-    st.write(f"🤖 AI Response: {reply}")
-
-    if st.checkbox("Use cloned voice to reply"):
-        if st.session_state.clone_path:
-            clone_audio = clone_and_generate_text(reply, st.session_state.clone_path)
-            st.audio(clone_audio)
-        else:
-            st.warning("No clone voice found. Please upload or record one.")
+    st.success(f"🤖 AI says: {reply}")
+
+    # Step 3: Speak back in your clone voice
+    if st.session_state.clone_path:
+        st.info("Generating voice reply using your cloned voice...")
+        voice_output_path = clone_and_generate_text(reply, st.session_state.clone_path)
+        st.audio(voice_output_path)
+    else:
+        st.warning("Please upload your clone voice first in the sidebar.")