Update app.py

app.py

@@ -1,84 +1,51 @@
 import gradio as gr
 from transformers import pipeline
-import tempfile
-import os
 
-# --- Core AI Model Setup ---
 pipe = pipeline("text-generation", model="Hulk810154/Kai", trust_remote_code=True)
 
-# --- Core Chat Functionality ---
-def chat_fn(message, history=None):
-    # AGI-style persistent conversation
-    if history is None:
-        history = []
+def chat_fn(message, history):
     prompt = ""
     for user, bot in history:
         prompt += f"User: {user}\nAI: {bot}\n"
     prompt += f"User: {message}\nAI:"
-    result = pipe(prompt, max_length=256, do_sample=True)
-    reply = result[0]["generated_text"][len(prompt):].strip()
+    output = pipe(prompt, max_length=256, do_sample=True)
+    reply = output[0]["generated_text"][len(prompt):].strip()
     return reply
 
-# --- Voice Input (Speech-to-Text) ---
-def voice_to_text(audio):
-    # Uses gradio's built-in whisper-based speech recognition
-    return audio
-
-# --- Text-to-Speech (TTS) ---
-def tts_fn(text):
-    return text  # Gradio auto-handles text-to-speech with the 'audio' output type
-
-# --- File Upload Handler ---
 def file_upload(files):
-    file_info = []
-    for file in files:
-        file_info.append(f"Uploaded: {os.path.basename(file.name)} ({file.size//1024} KB)")
-    return "\n".join(file_info)
-
-# --- Screenshot Upload (Pseudo-Screen-Share) ---
-def screenshot_upload(img):
-    return "Screenshot received!"
+    return f"Received {len(files)} files."
 
-# --- AGI Customization ---
-with gr.Blocks(theme=gr.themes.Base()) as demo:
-    gr.Markdown(
-        "# 🧠 **Kai AGI Chat**\n"
-        "Live voice + text, upload files/images, and more. Powered by Hugging Face."
-    )
+with gr.Blocks() as demo:
+    gr.Markdown("# Kai AGI – Text, Voice, & Multimodal Chat\n**Text, speak, or upload files/images. True AGI context.**")
     with gr.Row():
-        persona = gr.Textbox(label="Agent Persona Prompt (Optional)", placeholder="e.g., Act as a business advisor...", interactive=True)
-        session_history = gr.State([])
-
-    with gr.Tab("💬 Chat"):
-        chat = gr.ChatInterface(
-            fn=chat_fn,
-            additional_inputs=[persona],
-            text_to_speech="auto",
-            speech_to_text="whisper",
-            retry_btn=None,
-            undo_btn=None,
-            title="Kai – AGI Chat",
-            description="Text, talk, and upload files. True multimodal AGI mode. (Voice features require browser permission.)",
-        )
-
-    with gr.Tab("🎙️ Voice Chat"):
-        mic = gr.Audio(source="microphone", type="filepath", label="Speak here")
-        voice_output = gr.Textbox(label="Recognized Text")
-        tts_button = gr.Button("Text-to-Speech")
-        tts_audio = gr.Audio(label="AI Speaks")
-        mic.change(lambda x: pipe(x, max_length=256, do_sample=True)[0]["generated_text"], mic, voice_output)
-        tts_button.click(tts_fn, inputs=voice_output, outputs=tts_audio)
-
-    with gr.Tab("📤 Upload File / Image"):
-        file_input = gr.File(file_count="multiple")
-        file_status = gr.Textbox(label="File status")
-        file_input.upload(file_upload, inputs=file_input, outputs=file_status)
-
-    with gr.Tab("🖥️ Share Screenshot"):
-        screenshot = gr.Image(label="Upload Screenshot")
-        screenshot_status = gr.Textbox(label="Status")
-        screenshot.upload(screenshot_upload, inputs=screenshot, outputs=screenshot_status)
-
-    gr.Markdown("**Session Memory:** Your conversation will persist during this session for AGI-level context.")
+        with gr.Column(scale=2):
+            chatbot = gr.Chatbot()
+            msg = gr.Textbox(label="Type a message or use voice...", placeholder="Type here...", scale=2)
+            voice_input = gr.Audio(source="microphone", type="filepath", label="🎤 Voice input (click mic)")
+            send_btn = gr.Button("Send", elem_id="send-btn")
+        with gr.Column(scale=1):
+            upload = gr.File(file_count="multiple", label="📎 Upload files/images")
+            upload_status = gr.Textbox(label="Upload status")
+    state = gr.State([])
+
+    def user_message(text, audio, history):
+        if text:
+            return text, history
+        if audio:
+            import whisper
+            model = whisper.load_model("tiny")
+            result = model.transcribe(audio)
+            return result["text"], history
+        return "", history
+
+    def respond(message, history):
+        reply = chat_fn(message, history)
+        history = history + [(message, reply)]
+        return history, history
+
+    send_btn.click(user_message, inputs=[msg, voice_input, state], outputs=[msg, state], queue=False).then(
+        respond, inputs=[msg, state], outputs=[chatbot, state]
+    )
+    upload.upload(file_upload, inputs=upload, outputs=upload_status)
 
 demo.launch()