Refactor for Headshort and Scene Generation using Instant-ID model hosted in Replicate

.env.example

@@ -0,0 +1,16 @@
+# ==========================================
+# FaceForge AI - Environment Configuration
+# Copy this file to .env and update values
+# ==========================================
+
+# Dev Mode (optional)
+# false = 5 generations/day (default)
+# true = 10 generations/day (for development/testing)
+DEV_MODE=false
+
+
+
+# Replicate API Token (required)
+# Get yours at: https://replicate.com/account/api-tokens
+REPLICATE_API_TOKEN=your_token_here
+

.gitignore

@@ -1 +1,4 @@
-roadmap/
+roadmap/
+output/
+prompts
+*.bak

README.md

@@ -1,67 +1,61 @@
----
-title: FaceForgeAI ZeroGPU
-emoji: 🐨
-colorFrom: pink
-colorTo: pink
-sdk: gradio
-sdk_version: 5.49.1
-app_file: app.py
-pinned: false
-license: mit
-short_description: FaceForgeAI_ZeroGPU
----
+# 🎨 FaceForge AI
+[![Hugging Face Space](https://img.shields.io/badge/🤗%20Open%20in-Hugging%20Face%20Space-yellow)](https://huggingface.co/spaces/VcRlAgent/FaceForgeAI)
 
-# 🎨 FaceForge AI – ZeroGPU Gradio Edition  
-[![Hugging Face Space](https://img.shields.io/badge/🤗%20Open%20in-Hugging%20Face%20Space-yellow)](https://huggingface.co/spaces/VcRlAgent/FaceForgeAI_ZeroGPU)
+**Author:** Vijay S. Chaudhari  
+**Runtime:** Gradio + Replicate API 🚀
 
-**Author:** Vijay S. Chaudhari
-**Runtime:** Hugging Face Spaces (ZeroGPU) 🚀  
- 
 ---
 
-## 🧠 Overview  
-**FaceForge AI** transforms your uploaded photo into professional-quality images powered by open-source generative AI:  
-- 🪄 **Background Remover** – clean gradient background for profile photos  
-- 🛂 **Passport Photo** – compliant 600×600 white-background image  (Requires picture with frontal face)
-- 🎭 **Stylized AI Avatar** – realistic yet personalized stylization  
+## 🧠 Overview
+**FaceForge AI** transforms your uploaded photo into professional-quality images powered by AI:
+- 🎭 **Background Remover** – clean background removal for any image
+- 💼 **Professional Headshots** – studio-quality portraits in multiple styles
+- 🌍 **Scene Changer** – place yourself in different environments with identity preservation
 
-This edition is optimized for **ZeroGPU Spaces** — efficient, on-demand GPU execution using CPU offload and attention slicing for lightweight inference.
+This edition uses **Replicate's InstantID** for high-quality identity-preserving transformations with configurable rate limiting for cost control.
 
 ---
 
-## ⚙️ Key Features  
-✅ **Three Modes**
-- **Background Remover** – removes background and enhances clarity  
-- **Passport** – white background, standard size  
-- **Avatar** – realistic stylization via Stable Diffusion Img2Img  
+## ✨ Key Features
+
+✅ **Three Generation Modes**
+- **Background Remover** – removes background using Rembg
+- **Professional Headshots** – Studio, Office, Premium Editorial styles
+- **Scene Changer** – Coastal Run, Urban Evening, Forest Trail environments
 
-✅ **User Control for Avatar Generation**
-- Preset prompt styles + custom text input  
-- Adjustable `strength` & `guidance_scale` sliders  
+✅ **User Controls**
+- Preset prompt styles via dropdown
+- Optional advanced settings (CFG, Steps, Denoise)
+- Real-time usage tracking
+
+✅ **Rate Limiting**
+- 5 generations/day (standard mode)
+- 10 generations/day (dev mode)
+- Midnight UTC reset
+- Device-based tracking
 
 ---
 
-## 🧩 Tech Stack  
+## 🧩 Tech Stack
 
 | Component | Purpose |
-|------------|----------|
+|-----------|---------|
 | **Python 3.10+** | Core runtime |
 | **Gradio 4.x** | Web UI framework |
-| **Stable Diffusion v1.5** | Img2Img stylization |
-| **GFPGAN 1.3.8** | Facial restoration |
-| **Real-ESRGAN 0.3.0** | Super-resolution |
+| **Replicate API** | InstantID model hosting |
+| **InstantID** | Identity-preserving image generation |
 | **Rembg 2.x** | Background removal |
-| **Pillow / OpenCV / Torch** | Image processing + GPU acceleration |
-
+| **PyYAML** | Configuration management |
+| **Pillow** | Image processing |
 
 ---
 
-## 🧰 Installation  
+## 🧰 Installation
 
 ### 1️⃣ Clone Repository
 ```bash
 git clone https://github.com/agentofAI/FaceForgeAI.git
-cd FaceForgeAI_ZeroGPU
+cd FaceForgeAI
 ```
 
 ### 2️⃣ Install Dependencies
@@ -69,7 +63,15 @@ cd FaceForgeAI_ZeroGPU
 pip install -r requirements.txt
 ```
 
-### 3️⃣ Run Locally
+### 3️⃣ Configure Environment
+```bash
+cp .env.example .env
+# Edit .env and add your REPLICATE_API_TOKEN
+```
+
+Get your Replicate API token at: https://replicate.com/account/api-tokens
+
+### 4️⃣ Run Locally
 ```bash
 python app.py
 ```
@@ -77,46 +79,132 @@ Open the local Gradio URL (typically http://127.0.0.1:7860) in your browser.
 
 ---
 
-## 🎨 Avatar Prompt Presets
+## 🎨 Style Presets
 
-| Style Label | Prompt |
-|--------------|--------|
-| 🎬 **Cinematic Portrait** | highly detailed, digital portrait, professional lighting, cinematic style, artistic AI avatar |
-| 🎨 **Stylized Realism** | stylized yet realistic portrait, balanced lighting, subtle gradient background, sharp focus on face |
-| 🏢 **Studio Professional** | studio portrait, even lighting, neutral background, realistic skin, confident pose |
-| 🤵 **Natural Headshot** | realistic professional headshot, soft studio lighting, neutral background, crisp details, natural skin tone |
+### Professional Headshots
 
----
+| Style | Description |
+|-------|-------------|
+| 📸 **Studio Headshot** | Professional editorial with neutral background, magazine quality |
+| 🏢 **Office Setting** | Subtle office background, premium business attire |
+| ✨ **Premium Editorial** | High-end minimal studio, composed presence |
 
-## 🧾 Model Credits
-| Model                     | Source / License                                                                                              |
-| ------------------------- | ------------------------------------------------------------------------------------------------------------- |
-| **Stable Diffusion v1.5** | [runwayml/stable-diffusion-v1-5](https://huggingface.co/runwayml/stable-diffusion-v1-5) (CompVis / Runway ML) |
-| **GFPGAN v1.3**           | [TencentARC/GFPGAN](https://github.com/TencentARC/GFPGAN) (MIT License)                                       |
-| **Real-ESRGAN x2Plus**    | [xinntao/Real-ESRGAN](https://github.com/xinntao/Real-ESRGAN) (BSD License)                                   |
-| **Rembg**                 | [danielgatis/rembg](https://github.com/danielgatis/rembg)                                                     |
-| **Gradio**                | [gradio-app/gradio](https://github.com/gradio-app/gradio)                                                     |
+### Scene Changer
+
+| Scene | Description |
+|-------|-------------|
+| 🏖️ **Coastal Run** | Ocean and horizon background, warm sunrise lighting |
+| 🌃 **Urban Evening** | City lights, soft ambient night lighting |
+| 🌲 **Forest Trail** | Trees and dirt path, diffused outdoor light |
 
 ---
 
-## 🧠 Project Structure
+## 📁 Project Structure
+
 ```
 faceforge-ai/
 │
-├── app.py                # Main application
-├── requirements.txt      # Dependencies
-├── assets/               # Optional screenshots and samples
-└── README.md             # Documentation
+├── app.py                  # Main Gradio application
+├── config.yaml             # Prompts and settings configuration
+├── rate_limiter.py         # Rate limiting logic
+├── replicate_handler.py    # Replicate API wrapper
+├── requirements.txt        # Python dependencies
+├── .env.example            # Environment variables template
+├── rate_limits.json        # Auto-generated usage tracking
+└── README.md               # Documentation
+```
+
+---
+
+## ⚙️ Configuration
+
+### Rate Limits (config.yaml)
+```yaml
+rate_limit:
+  default_daily_limit: 5      # Standard mode
+  dev_daily_limit: 10         # Dev mode
+  reset_timezone: "UTC"       # Reset at midnight UTC
+```
+
+### Dev Mode
+Enable higher rate limits in `.env`:
+```bash
+DEV_MODE=true
 ```
 
+### Add Custom Styles
+Edit `config.yaml` to add new headshot or scene styles:
+
+```yaml
+headshots:
+  "🎯 Your Style":
+    prompt: |
+      Your custom prompt here...
+      multiple lines supported
+    negative: "Things to avoid..."
+
+scenes:
+  "🎪 Your Scene":
+    prompt: "Scene description..."
+    negative: "Things to avoid..."
+```
+
+### Advanced Settings
+Toggle "Enable Advanced Settings" in UI to control:
+- **CFG Scale**: Prompt adherence (1.0-10.0)
+- **Steps**: Generation quality (20-50)
+- **Denoise**: Transformation strength (0.5-1.0)
+
+---
+
+## 📊 Rate Limiting
+
+- **Tracking**: Server-side session file per device
+- **Reset**: Midnight UTC daily
+- **Scope**: Shared across all generation functions
+- **Override**: Dev mode doubles the limit
+
+Device fingerprinting uses IP + User-Agent for consistent tracking.
+
+---
+
+## 🧾 Model Credits
+
+| Model | Source / License |
+|-------|------------------|
+| **InstantID** | [zsxkib/instant-id-basic](https://replicate.com/zsxkib/instant-id-basic) via Replicate |
+| **Rembg** | [danielgatis/rembg](https://github.com/danielgatis/rembg) (MIT License) |
+| **Gradio** | [gradio-app/gradio](https://github.com/gradio-app/gradio) (Apache 2.0) |
+
+---
+
+## 🐛 Troubleshooting
+
+**"REPLICATE_API_TOKEN not found"**
+- Verify `.env` file exists with valid token
+- Check token at https://replicate.com/account/api-tokens
+
+**"Daily limit reached"**
+- Wait for midnight UTC reset
+- Enable `DEV_MODE=true` for higher limits
+
+**Generation fails**
+- Verify image uploaded successfully
+- Check Replicate API status
+- Ensure internet connectivity
+
+**Rate limits not working**
+- Delete `rate_limits.json` to reset tracking
+- Verify server-side file permissions
+
 ---
 
 ## 📜 License
 
-This project is for educational and demonstration purposes.
+This project is for educational and demonstration purposes.  
 Each model used retains its original open-source license.
 
 ---
-## 👨‍💻 Author
 
-Vijay S. Chaudhari
+Author: Vijay S. Chaudhari
+© 2025 Vijay S. Chaudhari

app.py

@@ -1,180 +1,149 @@
-# ==========================================
-# FaceForge AI – ZeroGPU Gradio Version
-# Author: Vijay S. Chaudhari | 2025
-# ==========================================
-
 import gradio as gr
-import spaces
+import yaml
 import torch
-import cv2
-import numpy as np
-from PIL import Image, ImageEnhance, ImageOps
+from PIL import Image
 from rembg import remove
-from diffusers import StableDiffusionImg2ImgPipeline
-from accelerate import Accelerator
-import io
-
-
-import torchvision
-print("Printing Torch and TorchVision versions:")
-print(torch.__version__)
-print(torchvision.__version__)
-
-# GPU libraries
-from gfpgan import GFPGANer
-from basicsr.archs.rrdbnet_arch import RRDBNet
-from realesrgan import RealESRGANer
+from rate_limiter import RateLimiter
+from replicate_handler import ReplicateHandler
 
-accelerator = Accelerator()
 
 # ------------------------------------------
-# Model Loading (Outside GPU decorator)
+# Load Configuration
 # ------------------------------------------
 
-def load_models():
-    """Load models once at startup"""
-    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    
-    # RealESRGAN upsampler
-    model = RRDBNet(num_in_ch=3, num_out_ch=3, num_feat=64, num_block=23, num_grow_ch=32, scale=2)
-    upsampler = RealESRGANer(
-        scale=2,
-        model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth',
-        model=model,
-        tile=400,
-        tile_pad=10,
-        pre_pad=0,
-        half=True,
-        device=device
-    )
-    
-    # GFPGAN enhancer
-    face_enhancer = GFPGANer(
-        model_path='https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth',
-        upscale=2,
-        arch='clean',
-        channel_multiplier=2,
-        bg_upsampler=upsampler,
-        device=device
-    )
-    
-    # Stable Diffusion Img2Img pipeline (public model)
-    device = "cuda" if torch.cuda.is_available() else "cpu"
-    if device == "cuda":
-        sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
-            "runwayml/stable-diffusion-v1-5",
-            torch_dtype=torch.float16
-        ).to(device)
-
-        # Optimize for ZeroGPU memory    
-        sd_pipe.enable_attention_slicing()
-        sd_pipe.enable_model_cpu_offload()
-
-    else:
-        sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
-            "runwayml/stable-diffusion-v1-5").to(device)  
-        sd_pipe.to(device)
+with open("config.yaml", "r") as f:
+    config = yaml.safe_load(f)
 
-    return face_enhancer, sd_pipe
+# Initialize modules
+rate_limiter = RateLimiter(
+    session_file=config["rate_limit"]["session_file"],
+    daily_limit=config["rate_limit"]["default_daily_limit"],
+    dev_daily_limit=config["rate_limit"]["dev_daily_limit"]
+)
 
-# Load models globally
-face_enhancer, sd_pipe = load_models()
+replicate_handler = ReplicateHandler(
+    model=config["replicate"]["model"],
+    default_settings=config["replicate"]["default_settings"]
+)
 
 # ------------------------------------------
-# GPU-Accelerated Functions
+# Core Functions
 # ------------------------------------------
 
-@spaces.GPU
-def enhance_face(img: Image.Image) -> Image.Image:
-    """Enhance face using GFPGAN (GPU)"""
-    img_cv = cv2.cvtColor(np.array(img.convert('RGB')), cv2.COLOR_RGB2BGR)
+def remove_background(img: Image.Image) -> Image.Image:
+    """Remove background from image"""
+    if img is None:
+        raise gr.Error("Please upload an image first")
+    return remove(img)
+
+
+def generate_headshot(
+    img: Image.Image,
+    headshot_style: str,
+    use_advanced: bool,
+    cfg: float,
+    steps: int,
+    denoise: float,
+    request: gr.Request
+) -> Image.Image:
+    """Generate professional headshot with selected style"""
     
-    with torch.no_grad():
-        _, _, restored_img = face_enhancer.enhance(
-            img_cv,
-            has_aligned=False,
-            only_center_face=False,
-            paste_back=True,
-            weight=0.5
-        )
+    # Check rate limit
+    allowed, remaining, reset_time = rate_limiter.check_limit(request)
+    if not allowed:
+        raise gr.Error(rate_limiter.get_limit_message(remaining, reset_time))
     
-    restored_img = cv2.cvtColor(restored_img, cv2.COLOR_BGR2RGB)
-    return Image.fromarray(restored_img)
-
-# ------------------------------------------
-# Image Processing Functions
-# ------------------------------------------
-
-def enhance_image(img: Image.Image) -> Image.Image:
-    """Basic enhancement"""
-    img = ImageEnhance.Contrast(img).enhance(1.15)
-    img = ImageEnhance.Sharpness(img).enhance(1.1)
-    return img
-
-@spaces.GPU
-def create_headshot(img: Image.Image) -> Image.Image:
-    """Professional headshot with gradient background"""
-    # Enhance face
-    img_enhanced = enhance_face(img)
+    if img is None:
+        raise gr.Error("Please upload an image first")
     
-    # Remove background
-    img_no_bg = remove(img_enhanced)
+    # Get prompt from config
+    style_config = config["headshots"][headshot_style]
+    prompt = style_config["prompt"]
+    negative_prompt = style_config["negative"]
     
-    # Gradient background
-    bg = Image.new("RGB", img_no_bg.size, (200, 210, 230))
-    if img_no_bg.mode == 'RGBA':
-        bg.paste(img_no_bg, mask=img_no_bg.split()[3])
+    # Prepare settings
+    custom_settings = None
+    if use_advanced:
+        custom_settings = {
+            "cfg": cfg,
+            "steps": steps,
+            "denoise": denoise
+        }
     
-    return enhance_image(bg)
-
-@spaces.GPU
-def create_passport(img: Image.Image) -> Image.Image:
-    """Passport photo with white background"""
-    # Enhance face
-    img_enhanced = enhance_face(img)
+    # Generate
+    try:
+        result = replicate_handler.generate(
+            input_image=img,
+            prompt=prompt,
+            negative_prompt=negative_prompt,
+            custom_settings=custom_settings
+        )
+        
+        # Increment usage
+        rate_limiter.increment(request)
+        
+        return result
+        
+    except Exception as e:
+        raise gr.Error(f"Generation failed: {str(e)}")
+
+
+def generate_scene(
+    img: Image.Image,
+    scene_style: str,
+    use_advanced: bool,
+    cfg: float,
+    steps: int,
+    denoise: float,
+    request: gr.Request
+) -> Image.Image:
+    """Generate scene change with selected style"""
     
-    # Remove background
-    img_no_bg = remove(img_enhanced)
+    # Check rate limit
+    allowed, remaining, reset_time = rate_limiter.check_limit(request)
+    if not allowed:
+        raise gr.Error(rate_limiter.get_limit_message(remaining, reset_time))
     
-    # White background (600x600)
-    bg = Image.new("RGB", (600, 600), (255, 255, 255))
-    img_no_bg.thumbnail((550, 550), Image.Resampling.LANCZOS)
-    offset = ((600 - img_no_bg.width) // 2, (600 - img_no_bg.height) // 2)
+    if img is None:
+        raise gr.Error("Please upload an image first")
     
-    if img_no_bg.mode == 'RGBA':
-        bg.paste(img_no_bg, offset, mask=img_no_bg.split()[3])
+    # Get prompt from config
+    style_config = config["scenes"][scene_style]
+    prompt = style_config["prompt"]
+    negative_prompt = style_config["negative"]
     
-    return bg
-
-@spaces.GPU
-def create_avatar(img: Image.Image, prompt: str, strength: float, guidance_scale: float) -> Image.Image:
-    """Stylized AI avatar using Stable Diffusion Img2Img with user inputs"""
-    # Enhance face
-    img_enhanced = enhance_face(img)
+    # Prepare settings
+    custom_settings = None
+    if use_advanced:
+        custom_settings = {
+            "cfg": cfg,
+            "steps": steps,
+            "denoise": denoise
+        }
     
-    # Resize for SD (512x512)
-    img_resized = img_enhanced.convert("RGB").resize((512, 512))
-
-    # Stylize with SD prompt. We are selecting these from UI now.
-    #prompt = "highly detailed, digital portrait, professional lighting, cinematic style, artistic AI avatar"
-    #prompt = "stylized yet realistic portrait, balanced lighting, subtle gradient background, sharp focus on face"
-    #prompt = "studio portrait, even lighting, neutral background, realistic skin, confident pose"
-    #prompt = "realistic professional headshot, soft studio lighting, neutral background, crisp details, natural skin tone"    
+    # Generate
+    try:
+        result = replicate_handler.generate(
+            input_image=img,
+            prompt=prompt,
+            negative_prompt=negative_prompt,
+            custom_settings=custom_settings
+        )
+        
+        # Increment usage
+        rate_limiter.increment(request)
+        
+        return result
+        
+    except Exception as e:
+        raise gr.Error(f"Generation failed: {str(e)}")
 
-    with torch.autocast("cuda"):
-        result = sd_pipe(prompt=prompt, image=img_resized, strength=strength, guidance_scale=guidance_scale)
 
-    avatar = enhance_face(result.images[0])
-    
-    return avatar
+def get_rate_limit_status(request: gr.Request) -> str:
+    """Get current rate limit status"""
+    allowed, remaining, reset_time = rate_limiter.check_limit(request)
+    return rate_limiter.get_limit_message(remaining, reset_time)
 
-@spaces.GPU
-def process_all(img: Image.Image):
-    """Process all three types at once"""
-    headshot = create_headshot(img)
-    passport = create_passport(img)
-    avatar = create_avatar(img)
-    return headshot, passport, avatar
 
 # ------------------------------------------
 # Gradio Interface
@@ -184,98 +153,143 @@ with gr.Blocks(theme=gr.themes.Soft(), title="FaceForge AI") as demo:
     gr.Markdown(
         """
         # 🎨 FaceForge AI
-        ### GPU-Accelerated Professional Headshot & Avatar Generator  
-        Upload your photo and choose or customize how your AI avatar is generated.
+        ### AI-Powered Professional Photo Generator
+        Upload your photo and transform it into professional headshots or place yourself in different scenes.
         """
     )
-
-    # --- Define a mapping: Short Label -> Full Prompt Text ---
-    PROMPT_MAP = {
-        "🎬 Cinematic Portrait": "highly detailed, digital portrait, professional lighting, cinematic style, artistic AI avatar",
-        "🎨 Stylized Realism": "stylized yet realistic portrait, balanced lighting, subtle gradient background, sharp focus on face",
-        "🏢 Studio Professional": "studio portrait, even lighting, neutral background, realistic skin, confident pose",
-        "🤵 Natural Headshot": "realistic professional headshot, soft studio lighting, neutral background, crisp details, natural skin tone"
-    }
-
+    
+    # Rate limit status
+    rate_status = gr.Textbox(
+        label="📊 Usage Status",
+        interactive=False,
+        value="Loading..."
+    )
+    
     with gr.Row():
+        # Left Column - Input
         with gr.Column(scale=1):
             input_image = gr.Image(type="pil", label="📷 Upload Your Photo")
-
-            gr.Markdown("### ⚙️ Avatar Generation Settings")
-
-            # Dropdown shows short labels only
-            preset_prompt = gr.Dropdown(
-                label="🎨 Choose Avatar Style Preset",
-                choices=list(PROMPT_MAP.keys()),
-                value="🤵 Natural Headshot"
+            
+            gr.Markdown("### ⚙️ Advanced Settings (Optional)")
+            
+            use_advanced = gr.Checkbox(
+                label="Enable Advanced Settings",
+                value=False
             )
-
-            # Optional custom prompt box for flexibility
-            custom_prompt = gr.Textbox(
-                label="✏️ Custom Prompt (optional)",
-                placeholder="Enter your own prompt or leave blank to use preset...",
-                lines=2
+            
+            with gr.Group(visible=True) as advanced_group:
+                cfg_slider = gr.Slider(
+                    label="🎛️ CFG Scale",
+                    minimum=config["replicate"]["advanced_ranges"]["cfg"]["min"],
+                    maximum=config["replicate"]["advanced_ranges"]["cfg"]["max"],
+                    value=config["replicate"]["default_settings"]["cfg"],
+                    step=config["replicate"]["advanced_ranges"]["cfg"]["step"]
+                )
+                
+                steps_slider = gr.Slider(
+                    label="🔄 Steps",
+                    minimum=config["replicate"]["advanced_ranges"]["steps"]["min"],
+                    maximum=config["replicate"]["advanced_ranges"]["steps"]["max"],
+                    value=config["replicate"]["default_settings"]["steps"],
+                    step=config["replicate"]["advanced_ranges"]["steps"]["step"]
+                )
+                
+                denoise_slider = gr.Slider(
+                    label="✨ Denoise",
+                    minimum=config["replicate"]["advanced_ranges"]["denoise"]["min"],
+                    maximum=config["replicate"]["advanced_ranges"]["denoise"]["max"],
+                    value=config["replicate"]["default_settings"]["denoise"],
+                    step=config["replicate"]["advanced_ranges"]["denoise"]["step"]
+                )
+            
+            # Toggle visibility of advanced settings
+            use_advanced.change(
+                fn=lambda x: gr.update(visible=x),
+                inputs=[use_advanced],
+                outputs=[advanced_group]
             )
-
-            strength_slider = gr.Slider(
-                label="🎛️ Style Strength (0.0 = keep original, 1.0 = full restyle)",
-                minimum=0.1,
-                maximum=1.0,
-                value=0.45,
-                step=0.05
+        
+        # Right Column - Outputs
+        with gr.Column(scale=1):
+            
+            # Background Remover Section
+            gr.Markdown("### 🎭 Background Remover")
+            output_bg_removed = gr.Image(label="Background Removed", type="pil")
+            btn_bg_remove = gr.Button("🗑️ Remove Background", variant="secondary")
+            
+            gr.Markdown("---")
+            
+            # Professional Headshots Section
+            gr.Markdown("### 💼 Professional Headshots")
+            headshot_style = gr.Dropdown(
+                label="Choose Headshot Style",
+                choices=list(config["headshots"].keys()),
+                value=list(config["headshots"].keys())[0]
             )
-
-            guidance_slider = gr.Slider(
-                label="🎯 Prompt Guidance Scale (higher = more prompt influence)",
-                minimum=1.0,
-                maximum=10.0,
-                value=5.5,
-                step=0.5
+            output_headshot = gr.Image(label="Professional Headshot", type="pil")
+            btn_headshot = gr.Button("📸 Generate Headshot", variant="primary")
+            
+            gr.Markdown("---")
+            
+            # Scene Changer Section
+            gr.Markdown("### 🌍 Scene Changer")
+            scene_style = gr.Dropdown(
+                label="Choose Scene",
+                choices=list(config["scenes"].keys()),
+                value=list(config["scenes"].keys())[0]
             )
-
-        with gr.Column(scale=1):
-            gr.Markdown("### Results")
-
-            # --- Independent Outputs & Buttons ---
-            output_headshot = gr.Image(label="💼 Professional Headshot", type="pil")
-            btn_headshot = gr.Button("📸 Generate Headshot", variant="secondary")
-
-            output_passport = gr.Image(label="🛂 Passport Photo", type="pil")
-            btn_passport = gr.Button("🪪 Generate Passport", variant="secondary")
-
-            output_avatar = gr.Image(label="🎭 AI Avatar", type="pil")
-            btn_avatar = gr.Button("✨ Generate Avatar", variant="primary")
-
-    # --- Functions for individual generations ---
-    def run_headshot(img):
-        return create_headshot(img)
-
-    def run_passport(img):
-        return create_passport(img)
-
-    def run_avatar(img, preset_label, custom, strength, guidance):
-        final_prompt = custom.strip() if custom and custom.strip() != "" else PROMPT_MAP[preset_label]
-        return create_avatar(img, final_prompt, strength, guidance)
-
-    # --- Button actions ---
-    btn_headshot.click(fn=run_headshot, inputs=[input_image], outputs=[output_headshot])
-    btn_passport.click(fn=run_passport, inputs=[input_image], outputs=[output_passport])
-    btn_avatar.click(
-        fn=run_avatar,
-        inputs=[input_image, preset_prompt, custom_prompt, strength_slider, guidance_slider],
-        outputs=[output_avatar]
+            output_scene = gr.Image(label="Scene Changed", type="pil")
+            btn_scene = gr.Button("🎬 Change Scene", variant="primary")
+    
+    # Button Actions
+    btn_bg_remove.click(
+        fn=remove_background,
+        inputs=[input_image],
+        outputs=[output_bg_removed]
+    )
+    
+    btn_headshot.click(
+        fn=generate_headshot,
+        inputs=[
+            input_image,
+            headshot_style,
+            use_advanced,
+            cfg_slider,
+            steps_slider,
+            denoise_slider
+        ],
+        outputs=[output_headshot]
     )
     
+    btn_scene.click(
+        fn=generate_scene,
+        inputs=[
+            input_image,
+            scene_style,
+            use_advanced,
+            cfg_slider,
+            steps_slider,
+            denoise_slider
+        ],
+        outputs=[output_scene]
+    )
+    
+    # Update rate limit status on load and after each generation
+    demo.load(fn=get_rate_limit_status, inputs=None, outputs=[rate_status])
+    btn_headshot.click(fn=get_rate_limit_status, inputs=None, outputs=[rate_status])
+    btn_scene.click(fn=get_rate_limit_status, inputs=None, outputs=[rate_status])
+    
     gr.Markdown(
         """
         ---
-        ### Features
-        - 💼 **Professional Headshots**: Perfect for LinkedIn and business profiles
-        - 🛂 **Passport Photos**: Standard 600x600px with white background
-        - 🎭 **AI Avatars**: Stylized versions for social media (Work In Progress - Identity Retention)
-        - ⚡ **GPU-Accelerated**: Fast processing with GFPGAN enhancement
+        ### ✨ Features
+        - 🎭 **Background Remover**: Clean background removal for any image
+        - 💼 **Professional Headshots**: Studio-quality headshots with multiple styles
+        - 🌍 **Scene Changer**: Place yourself in different environments
+        - ⚡ **AI-Powered**: InstantID technology for identity preservation
+        - 🔒 **Rate Limited**: Fair usage with daily limits
         
-        © 2025 Vijay S. Chaudhari | Powered by ZeroGPU 🚀
+        © 2025 Vijay S. Chaudhari | Powered by Replicate 🚀
         """
     )
 

app.py.wip_avatar

@@ -1,423 +0,0 @@
-# ==========================================
-# FaceForge AI – ZeroGPU Gradio Version
-# Author: Vijay S. Chaudhari | 2025
-# ==========================================
-
-import importlib.util
-import gradio as gr
-import spaces
-import torch
-import cv2
-import numpy as np
-from pathlib import Path
-
-from PIL import Image, ImageEnhance, ImageOps
-from rembg import remove
-from diffusers import StableDiffusionImg2ImgPipeline
-from diffusers import StableDiffusionXLPipeline
-import io
-import os, sys, subprocess, warnings, logging
-
-warnings.filterwarnings("ignore", category=UserWarning)
-logging.getLogger("onnxruntime").setLevel(logging.ERROR)
-os.environ["CUDA_VISIBLE_DEVICES"] = ""
-
-# --- Ensure InstantID is available ---
-if not Path("instantid").exists():
-    print("🔄 Cloning InstantID repository...")
-    subprocess.run(["git", "clone", "--depth", "1", "https://github.com/InstantID/InstantID.git", "instantid"],check=True)
-
-repo_root = Path("instantid").resolve()
-
-# 🧭 Search for a pipeline file that matches *instantid*.py under the repo
-candidates = list(repo_root.rglob("pipeline*instantid*.py"))
-if not candidates:
-    # Fallback common names across commits
-    fallback_names = [
-        "pipelines/pipeline_instantid.py",
-        "pipelines/pipeline_stable_diffusion_instantid.py",
-        "pipelines/pipeline_stable_diffusion_xl_instantid.py",
-    ]
-    for name in fallback_names:
-        p = repo_root / name
-        if p.exists():
-            candidates = [p]
-            break
-
-if not candidates:
-    raise FileNotFoundError(
-        "Could not locate an InstantID pipeline file under ./instantid. "
-        "Repo layout may have changed. Please check the repo structure."
-    )
-
-pipeline_file = candidates[0]
-print(f"✅ Using InstantID pipeline file: {pipeline_file.relative_to(repo_root)}")
-
-# 🪄 Import the pipeline module by file path (no package needed)
-spec = importlib.util.spec_from_file_location("instantid_pipeline", str(pipeline_file))
-instantid_mod = importlib.util.module_from_spec(spec)
-spec.loader.exec_module(instantid_mod)  # type: ignore
-
-# 🔎 Pick a pipeline class that looks like an InstantID Pipeline
-InstantIDPipeline = None
-for attr in dir(instantid_mod):
-    if "InstantID" in attr and "Pipeline" in attr:
-        InstantIDPipeline = getattr(instantid_mod, attr)
-        break
-
-if InstantIDPipeline is None:
-    # Helpful diagnostics
-    print("Available names in module:", [a for a in dir(instantid_mod) if "Pipeline" in a])
-    raise ImportError(
-        "Could not find an InstantID pipeline class. "
-        "Looked for a class name containing both 'InstantID' and 'Pipeline'."
-    )
-
-print(f"✅ Imported pipeline class: {InstantIDPipeline.__name__}")
-
-'''
-if os.path.exists("InstantID") and not os.path.exists("instantid"):
-    os.rename("InstantID", "instantid")    
-
-instantid_path = os.path.abspath("instantid")
-sys.path.append(instantid_path)
-sys.path.append(os.path.join(instantid_path, "pipelines"))
-
-#sys.path.append(os.path.abspath("instantid"))
-#sys.path.insert(0, os.path.join(os.getcwd(), 'InstantID'))
-try:
-    from pipelines.pipeline_instantid import InstantIDPipeline
-    print("✅ InstantIDPipeline imported successfully.")
-except Exception as e:
-    print("⚠️ Failed to import InstantIDPipeline:", e)
-    InstantIDPipeline = None  # graceful fallback
-'''
-
-import torchvision
-print("Printing Torch and TorchVision versions:")
-print(torch.__version__)
-print(torchvision.__version__)
-
-# GPU libraries
-from gfpgan import GFPGANer
-from basicsr.archs.rrdbnet_arch import RRDBNet
-from realesrgan import RealESRGANer
-
-# ------------------------------------------
-# Model Loading (Outside GPU decorator)
-# ------------------------------------------
-
-def load_models():
-    """Load models once at startup"""
-    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    
-    # RealESRGAN upsampler
-    model = RRDBNet(num_in_ch=3, num_out_ch=3, num_feat=64, num_block=23, num_grow_ch=32, scale=2)
-    upsampler = RealESRGANer(
-        scale=2,
-        model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth',
-        model=model,
-        tile=400,
-        tile_pad=10,
-        pre_pad=0,
-        half=True,
-        device=device
-    )
-    
-    # GFPGAN enhancer
-    face_enhancer = GFPGANer(
-        model_path='https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth',
-        upscale=2,
-        arch='clean',
-        channel_multiplier=2,
-        bg_upsampler=upsampler,
-        device=device
-    )
-    
-    # Stable Diffusion Img2Img pipeline (public model)
-    sd_pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
-        "runwayml/stable-diffusion-v1-5",
-        torch_dtype=torch.float16
-    ).to(device)
-
-    # Optimize for ZeroGPU memory
-    sd_pipe.enable_attention_slicing()
-    sd_pipe.enable_model_cpu_offload()
-
-    return face_enhancer, sd_pipe
-
-# Load models globally
-face_enhancer, sd_pipe = load_models()
-
-# ------------------------------------------
-# GPU-Accelerated Functions
-# ------------------------------------------
-
-@spaces.GPU
-def enhance_face(img: Image.Image) -> Image.Image:
-    """Enhance face using GFPGAN (GPU)"""
-    img_cv = cv2.cvtColor(np.array(img.convert('RGB')), cv2.COLOR_RGB2BGR)
-    
-    with torch.no_grad():
-        _, _, restored_img = face_enhancer.enhance(
-            img_cv,
-            has_aligned=False,
-            only_center_face=False,
-            paste_back=True,
-            weight=0.5
-        )
-    
-    restored_img = cv2.cvtColor(restored_img, cv2.COLOR_BGR2RGB)
-    return Image.fromarray(restored_img)
-
-# ------------------------------------------
-# Image Processing Functions
-# ------------------------------------------
-
-def enhance_image(img: Image.Image) -> Image.Image:
-    """Basic enhancement"""
-    img = ImageEnhance.Contrast(img).enhance(1.15)
-    img = ImageEnhance.Sharpness(img).enhance(1.1)
-    return img
-
-@spaces.GPU
-def create_headshot(img: Image.Image) -> Image.Image:
-    """Professional headshot with gradient background"""
-    # Enhance face
-    img_enhanced = enhance_face(img)
-    
-    # Remove background
-    img_no_bg = remove(img_enhanced)
-    
-    # Gradient background
-    bg = Image.new("RGB", img_no_bg.size, (200, 210, 230))
-    if img_no_bg.mode == 'RGBA':
-        bg.paste(img_no_bg, mask=img_no_bg.split()[3])
-    
-    return enhance_image(bg)
-
-@spaces.GPU
-def create_passport(img: Image.Image) -> Image.Image:
-    """Passport photo with white background"""
-    # Enhance face
-    img_enhanced = enhance_face(img)
-    
-    # Remove background
-    img_no_bg = remove(img_enhanced)
-    
-    # White background (600x600)
-    bg = Image.new("RGB", (600, 600), (255, 255, 255))
-    img_no_bg.thumbnail((550, 550), Image.Resampling.LANCZOS)
-    offset = ((600 - img_no_bg.width) // 2, (600 - img_no_bg.height) // 2)
-    
-    if img_no_bg.mode == 'RGBA':
-        bg.paste(img_no_bg, offset, mask=img_no_bg.split()[3])
-    
-    return bg
-
-@spaces.GPU
-def create_avatar(img: Image.Image, prompt: str, strength: float, guidance_scale: float) -> Image.Image:
-    """
-    Create a stylized AI avatar while preserving facial identity using InstantID.
-    Retains core facial features, skin tone, and expressions of the input photo.
-    """
-
-    # Stylize with SD prompt. We are selecting these from UI now.
-    #prompt = "highly detailed, digital portrait, professional lighting, cinematic style, artistic AI avatar"
-    #prompt = "stylized yet realistic portrait, balanced lighting, subtle gradient background, sharp focus on face"
-    #prompt = "studio portrait, even lighting, neutral background, realistic skin, confident pose"
-    #prompt = "realistic professional headshot, soft studio lighting, neutral background, crisp details, natural skin tone"    
-
-    # --- Convert input ---
-    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-    img = img.convert("RGB").resize((512, 512), Image.Resampling.LANCZOS)
-
-    # --- Step 1: Load InstantID + SDXL pipeline ---
-    pipe = StableDiffusionXLPipeline.from_pretrained(
-        "stabilityai/stable-diffusion-xl-base-1.0",
-        torch_dtype=torch.float16
-    ).to(device)
-
-    instantid = InstantIDPipeline.from_pretrained("InstantID/InstantID", torch_dtype=torch.float16,)
-    pipe.to("cuda" if torch.cuda.is_available() else "cpu")
-    #pipe.load_ip_adapter(instantid)
-
-    # --- Step 2: Optimize for ZeroGPU memory ---
-    pipe.enable_attention_slicing()
-    pipe.enable_model_cpu_offload()
-
-    # --- Step 3: Prepare conditioning (face embedding) ---
-    np_img = np.array(img)
-    bgr_img = cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR)
-    face_emb = instantid.extract_face_embedding(bgr_img)  # key step: ID embedding guidance
-
-    # --- Step 4: Stylized generation ---
-    gen = pipe.generate_with_identity(
-        image=img,
-        face_embedding=face_emb,
-        prompt=(
-            prompt
-            + ", portrait of the same person, consistent identity, detailed lighting, "
-              "highly realistic skin texture, cinematic color tones"
-        ),
-        strength=float(strength),
-        guidance_scale=float(guidance_scale),
-        num_inference_steps=30
-    )
-
-    avatar = gen.images[0]
-
-    # --- Step 5 (Optional): Post-process with GFPGAN for crispness ---
-    try:
-        from gfpgan import GFPGANer
-        from realesrgan import RealESRGANer
-        from basicsr.archs.rrdbnet_arch import RRDBNet
-
-        model = RRDBNet(num_in_ch=3, num_out_ch=3, num_feat=64,
-                        num_block=23, num_grow_ch=32, scale=2)
-        upsampler = RealESRGANer(
-            scale=2,
-            model_path='https://github.com/xinntao/Real-ESRGAN/releases/download/v0.2.1/RealESRGAN_x2plus.pth',
-            model=model,
-            tile=400,
-            tile_pad=10,
-            pre_pad=0,
-            half=True,
-            device=device
-        )
-        face_enhancer = GFPGANer(
-            model_path='https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.3.pth',
-            upscale=1,
-            arch='clean',
-            channel_multiplier=2,
-            bg_upsampler=upsampler,
-            device=device
-        )
-
-        img_cv = cv2.cvtColor(np.array(avatar), cv2.COLOR_RGB2BGR)
-        _, _, restored_img = face_enhancer.enhance(
-            img_cv, has_aligned=False, only_center_face=False,
-            paste_back=True, weight=0.4
-        )
-        avatar = Image.fromarray(cv2.cvtColor(restored_img, cv2.COLOR_BGR2RGB))
-    except Exception as e:
-        print(f"[WARN] GFPGAN post-process skipped: {e}")
-
-    return avatar
-
-
-@spaces.GPU
-def process_all(img: Image.Image):
-    """Process all three types at once"""
-    headshot = create_headshot(img)
-    passport = create_passport(img)
-    avatar = create_avatar(img)
-    return headshot, passport, avatar
-
-# ------------------------------------------
-# Gradio Interface
-# ------------------------------------------
-
-with gr.Blocks(theme=gr.themes.Soft(), title="FaceForge AI") as demo:
-    gr.Markdown(
-        """
-        # 🎨 FaceForge AI
-        ### GPU-Accelerated Professional Headshot & Avatar Generator  
-        Upload your photo and choose or customize how your AI avatar is generated.
-        """
-    )
-
-    # --- Define a mapping: Short Label -> Full Prompt Text ---
-    PROMPT_MAP = {
-        "🎬 Cinematic Portrait": "highly detailed, digital portrait, professional lighting, cinematic style, artistic AI avatar",
-        "🎨 Stylized Realism": "stylized yet realistic portrait, balanced lighting, subtle gradient background, sharp focus on face",
-        "🏢 Studio Professional": "studio portrait, even lighting, neutral background, realistic skin, confident pose",
-        "🤵 Natural Headshot": "realistic professional headshot, soft studio lighting, neutral background, crisp details, natural skin tone"
-    }
-
-    with gr.Row():
-        with gr.Column(scale=1):
-            input_image = gr.Image(type="pil", label="📷 Upload Your Photo")
-
-            gr.Markdown("### ⚙️ Avatar Generation Settings")
-
-            # Dropdown shows short labels only
-            preset_prompt = gr.Dropdown(
-                label="🎨 Choose Avatar Style Preset",
-                choices=list(PROMPT_MAP.keys()),
-                value="🤵 Natural Headshot"
-            )
-
-            # Optional custom prompt box for flexibility
-            custom_prompt = gr.Textbox(
-                label="✏️ Custom Prompt (optional)",
-                placeholder="Enter your own prompt or leave blank to use preset...",
-                lines=2
-            )
-
-            strength_slider = gr.Slider(
-                label="🎛️ Style Strength (0.0 = keep original, 1.0 = full restyle)",
-                minimum=0.1,
-                maximum=1.0,
-                value=0.45,
-                step=0.05
-            )
-
-            guidance_slider = gr.Slider(
-                label="🎯 Prompt Guidance Scale (higher = more prompt influence)",
-                minimum=1.0,
-                maximum=10.0,
-                value=5.5,
-                step=0.5
-            )
-
-        with gr.Column(scale=1):
-            gr.Markdown("### Results")
-
-            # --- Independent Outputs & Buttons ---
-            output_headshot = gr.Image(label="💼 Professional Headshot", type="pil")
-            btn_headshot = gr.Button("📸 Generate Headshot", variant="secondary")
-
-            output_passport = gr.Image(label="🛂 Passport Photo", type="pil")
-            btn_passport = gr.Button("🪪 Generate Passport", variant="secondary")
-
-            output_avatar = gr.Image(label="🎭 AI Avatar", type="pil")
-            btn_avatar = gr.Button("✨ Generate Avatar", variant="primary")
-
-    # --- Functions for individual generations ---
-    def run_headshot(img):
-        return create_headshot(img)
-
-    def run_passport(img):
-        return create_passport(img)
-
-    def run_avatar(img, preset_label, custom, strength, guidance):
-        final_prompt = custom.strip() if custom and custom.strip() != "" else PROMPT_MAP[preset_label]
-        return create_avatar(img, final_prompt, strength, guidance)
-
-    # --- Button actions ---
-    btn_headshot.click(fn=run_headshot, inputs=[input_image], outputs=[output_headshot])
-    btn_passport.click(fn=run_passport, inputs=[input_image], outputs=[output_passport])
-    btn_avatar.click(
-        fn=run_avatar,
-        inputs=[input_image, preset_prompt, custom_prompt, strength_slider, guidance_slider],
-        outputs=[output_avatar]
-    )
-    
-    gr.Markdown(
-        """
-        ---
-        ### Features
-        - 💼 **Professional Headshots**: Perfect for LinkedIn and business profiles
-        - 🛂 **Passport Photos**: Standard 600x600px with white background
-        - 🎭 **AI Avatars**: Stylized versions for social media
-        - ⚡ **GPU-Accelerated**: Fast processing with GFPGAN enhancement
-        
-        © 2025 Vijay S. Chaudhari | Powered by ZeroGPU 🚀
-        """
-    )
-
-# Launch
-if __name__ == "__main__":
-    demo.queue(max_size=20)
-    demo.launch()

config.yaml

@@ -0,0 +1,130 @@
+# FaceForge AI Configuration
+# Author: Vijay S. Chaudhari | 2025
+
+# Rate Limiting
+rate_limit:
+  default_daily_limit: 5
+  dev_daily_limit: 10
+  reset_timezone: "UTC"
+  session_file: "rate_limits.json"
+
+# Replicate Model
+replicate:
+  model: "zsxkib/instant-id-basic:beeceec1b60538a6b77c1549b8b1f91794b133c53443234feb8faa3c61ceab49"
+  
+  # Default settings (used when advanced options disabled)
+  default_settings:
+    cfg: 4.5
+    steps: 30
+    width: 1600
+    height: 1600
+    denoise: 1
+    scheduler: "karras"
+    sampler_name: "ddpm"
+    output_format: "webp"
+    output_quality: 90
+    instantid_weight: 1
+    instantid_start_at: 0
+    instantid_end_at: 1
+  
+  # Advanced settings ranges (for UI sliders when enabled)
+  advanced_ranges:
+    cfg:
+      min: 1.0
+      max: 10.0
+      step: 0.5
+    steps:
+      min: 20
+      max: 50
+      step: 5
+    denoise:
+      min: 0.5
+      max: 1.0
+      step: 0.1
+
+# Scene Change Prompts
+scenes:
+  "🏖️ Coastal Run":
+    prompt: |
+      Portrait of the same runner, black hoodie
+      running along a coastal path,
+      ocean and horizon in the background,
+      confident and relaxed expression,
+      warm sunrise lighting,
+      realistic lifestyle photography
+    negative: "NSFW, nudity, painting, drawing, illustration, glitch, deformed, mutated, cross-eyed, ugly, disfigured, wrinkles, open mouth, closed eyes, other characters"
+  
+  "🌃 Urban Evening":
+    prompt: |
+      Portrait of the same runner,
+      evening run in an urban environment,
+      city lights in the background,
+      confident relaxed smile,
+      soft ambient night lighting,
+      realistic low-light photography
+    negative: "NSFW, nudity, painting, drawing, illustration, glitch, deformed, mutated, cross-eyed, ugly, disfigured, wrinkles, open mouth, closed eyes, other characters"
+  
+  "🌲 Forest Trail":
+    prompt: |
+      Portrait of the same runner, identical person,
+      running on a forest trail,
+      trees and dirt path in the background,
+      relaxed expression,
+      diffused outdoor light,
+      realistic outdoor portrait
+    negative: "NSFW, nudity, painting, drawing, illustration, glitch, deformed, mutated, cross-eyed, ugly, disfigured, wrinkles, open mouth, closed eyes, other characters"
+
+# Professional Headshot Prompts
+headshots:
+  "📸 Studio Headshot":
+    prompt: |
+      Professional editorial headshot of the same person,
+      neutral studio background,
+      calm confident expression,
+      tailored professional attire,
+      soft diffused studio lighting,
+      natural skin texture,
+      realistic magazine photography
+    negative: |
+      cartoon, anime, illustration,
+      exaggerated features,
+      overly smooth skin,
+      plastic face,
+      dramatic shadows,
+      harsh lighting,
+      low quality
+  
+  "🏢 Office Setting":
+    prompt: |
+      High-end professional headshot of the same person,
+      subtle office setting in the background,
+      softly blurred professional indoor environment,
+      composed and confident presence,
+      premium business attire,
+      soft directional studio lighting,
+      realistic editorial photography
+    negative: |
+      artificial office,
+      fake interior,
+      overly sharp background,
+      glass walls,
+      corporate stock photo,
+      dramatic lighting,
+      cinematic contrast
+  
+  "✨ Premium Editorial":
+    prompt: |
+      High-end professional headshot of the same person,
+      minimal studio background,
+      composed and confident presence,
+      premium business attire,
+      soft directional studio lighting,
+      realistic editorial photography
+    negative: |
+      cartoon, anime, illustration,
+      exaggerated features,
+      overly smooth skin,
+      plastic face,
+      dramatic shadows,
+      harsh lighting,
+      low quality

env.example

@@ -0,0 +1,13 @@
+# ==========================================
+# FaceForge AI - Environment Configuration
+# Copy this file to .env and update values
+# ==========================================
+
+# Replicate API Token (required)
+# Get yours at: https://replicate.com/account/api-tokens
+REPLICATE_API_TOKEN=your_token_here
+
+# Dev Mode (optional)
+# false = 5 generations/day (default)
+# true = 10 generations/day (for development/testing)
+DEV_MODE=false

requirements.txt

@@ -1,46 +1,8 @@
-# ==========================================
-# FaceForge AI - ZeroGPU Gradio Requirements
-# ==========================================
-
-# Gradio & ZeroGPU (REQUIRED)
-gradio==4.16.0
-spaces==0.19.4
-
-# Image Processing
-Pillow==10.2.0
-opencv-python-headless==4.9.0.80
-numpy==1.26.3
-
-# Background Removal
-rembg[gpu]==2.0.57
-
-# Face Enhancement (GPU)
-gfpgan==1.3.8
-realesrgan==0.3.0
-basicsr==1.4.2
-
-# PyTorch (GPU) - Critical for CUDA
-torch==2.1.2
-torchvision==0.16.2
-#--extra-index-url https://download.pytorch.org/whl/cu118
-#--extra-index-url https://download.pytorch.org/whl/cu121
-
-# Stable Diffusion 
-diffusers>=0.20.0
-transformers==4.39.0
-accelerate==0.28.0
-
-# InstantID
-
-
-# Core Dependencies
-scipy==1.12.0
-tqdm==4.66.1
-facexlib==0.3.0
-yapf==0.40.2
-filterpy==1.4.5
-
-# Optional: For better performance
-# onnxruntime-gpu==1.17.0
-
-accelerate
+gradio>=4.0.0
+replicate>=0.20.0
+rembg>=2.0.0
+Pillow>=10.0.0
+PyYAML>=6.0
+torch>=2.0.0
+torchvision>=0.15.0
+requests>=2.31.0

src/replicateHandler.py

@@ -0,0 +1,90 @@
+import os
+import replicate
+from PIL import Image
+import io
+import base64
+import tempfile
+from typing import Optional
+
+
+class ReplicateHandler:
+    def __init__(self, model: str, default_settings: dict):
+        self.model = model
+        self.default_settings = default_settings
+        
+        # Verify API token
+        api_token = os.getenv("REPLICATE_API_TOKEN")
+        if not api_token:
+            raise ValueError("REPLICATE_API_TOKEN not found in environment variables")
+    
+    def _image_to_base64_url(self, image: Image.Image) -> str:
+        """Convert PIL Image to base64 data URL for Replicate"""
+        buffered = io.BytesIO()
+        image.save(buffered, format="PNG")
+        img_str = base64.b64encode(buffered.getvalue()).decode()
+        return f"data:image/png;base64,{img_str}"
+    
+    def _save_temp_image(self, image: Image.Image) -> str:
+        """Save image to temp file and return path"""
+        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".png")
+        image.save(temp_file.name, format="PNG")
+        return temp_file.name
+    
+    def generate(
+        self,
+        input_image: Image.Image,
+        prompt: str,
+        negative_prompt: str,
+        custom_settings: Optional[dict] = None
+    ) -> Image.Image:
+        """
+        Generate image using Replicate InstantID
+        
+        Args:
+            input_image: PIL Image
+            prompt: Positive prompt
+            negative_prompt: Negative prompt
+            custom_settings: Override default settings (cfg, steps, etc.)
+        
+        Returns:
+            Generated PIL Image
+        """
+        # Merge settings
+        settings = {**self.default_settings}
+        if custom_settings:
+            settings.update(custom_settings)
+        
+        # Save temp image and get file object
+        temp_path = self._save_temp_image(input_image)
+        
+        try:
+            # Prepare input
+            input_params = {
+                "image": open(temp_path, "rb"),
+                "prompt": prompt,
+                "negative_prompt": negative_prompt,
+                **settings
+            }
+            
+            # Run prediction (streaming)
+            output = replicate.run(self.model, input=input_params)
+            
+            # Get final image from iterator
+            result_url = None
+            for item in output:
+                result_url = item  # Last item is the final image URL
+            
+            if not result_url:
+                raise ValueError("No output received from Replicate")
+            
+            # Download and convert to PIL
+            import requests
+            response = requests.get(result_url)
+            result_image = Image.open(io.BytesIO(response.content))
+            
+            return result_image
+            
+        finally:
+            # Cleanup temp file
+            if os.path.exists(temp_path):
+                os.unlink(temp_path)