Update app.py

app.py

@@ -393,14 +393,51 @@ with gr.Blocks(title="Nano Vision Studio") as demo:
         # --- TAB 4 : GUIDE ---
         with gr.TabItem("📚 Guide"):
             gr.Markdown("""
-            ### 🍌 Quick Guide for Nano Vision Studio
-            
-            This interface is powered by **Gemini 3 Pro** and **Gemini 2.5 Flash**.
-            
-            * **Aspect Ratios**: Supports wide variety from 1:1 to 21:9.
-            * **Resolutions**: 1K (Default), 2K, 4K (Pro model only).
-            * **Multi-Image**: You can upload up to **14 images** in the Composition tab.
-            * **Chat**: Supports iterative editing ("Make it blue", "Add a cat").
+            # Comprehensive Guide
+
+            Welcome to the ultimate interface for **Nano Banana Pro** (Gemini 3 Pro) and **Nano Banana** (Gemini 2.5 Flash).
+
+            ## 🚀 Choose Your Model
+
+            | Feature | ⚡ Gemini 2.5 Flash (Nano Banana) | 🧠 Gemini 3 Pro (Nano Banana Pro) |
+            | :--- | :--- | :--- |
+            | **Best For** | Speed, High Volume, Prototyping | Professional Assets, Complex Logic, Text Rendering |
+            | **Resolution** | 1024x1024 (Native) | Up to **4K** (High Fidelity) |
+            | **Inputs** | Text + Images | Text + up to **14 Reference Images** |
+            | **Special** | Fast & Efficient | **Thinking Process**, **Search Grounding** |
+
+            ---
+
+            ## ✨ Advanced Capabilities Explained
+
+            ### 1. 🧠 The "Thinking" Process (Pro Only)
+            Nano Banana Pro doesn't just draw; it *thinks*. Before generating pixels, it reasons through your prompt to understand composition, lighting, and logic.
+            * **In this App:** Check the **"Thought Process"** accordion in the *Creation Studio* to read the model's internal monologue and see draft visualizations (Thought Images).
+
+            ### 2. 🌍 Search Grounding (Real-Time Data)
+            The model isn't stuck in the past. It can access **Google Search** to generate images based on live data.
+            * **Try this:** "Visualize the current weather forecast for Tokyo as a modern chart."
+            * **In this App:** Enable the **"Grounding"** checkbox. Sources will appear below the generated image.
+
+            ### 3. 🖼️ Advanced Composition (up to 14 Images)
+            While Flash handles fewer inputs, Pro can mix up to **14 images**!
+            * **Use Case:** Style transfer, maintaining character consistency, or complex collages.
+            * **How:** Use the **"Composition"** tab to upload multiple reference files.
+
+            ---
+
+            ## 💡 Prompting Masterclass
+
+            To get the best results, follow these professional tips:
+
+            * **Be Hyper-Specific:** Don't just say "a cat". Say *"A photorealistic close-up of a Siamese cat, golden hour lighting, captured with an 85mm lens"*.
+            * **Provide Context:** Explain the intent. *"Create a logo for a high-end minimalist skincare brand"*.
+            * **Positive Framing:** Instead of "no cars", describe the scene as *"an empty, deserted street"*.
+            * **Iterate with Chat:** Don't expect perfection on turn #1. Use the **Chat & Refinement** tab to say *"Make the lighting warmer"* or *"Add a wizard hat"*.
+
+            ## ⚡ Performance Tips
+            * **4K Generation:** Available only on Pro. It costs more but delivers stunning print-quality results.
+            * **Aspect Ratios:** We support everything from **1:1** to **21:9** (Cinematic).
             """)
 
 if __name__ == "__main__":