diff --git a/Caddyfile b/Caddyfile
index 379004e..10c40c7 100644
--- a/Caddyfile
+++ b/Caddyfile
@@ -1,15 +1,122 @@
-# Caddy 2 Configuration for AI Photo Edit
-# Simple version - frontend nginx handles API routing internally
+# Caddy 2 Configuration for EditmaskwithAI
+# ==========================================
+#
+# SETUP INSTRUCTIONS:
+# 1. Replace 'your-subdomain.yourdomain.com' with your actual domain
+# 2. Make sure DNS CNAME record points to your server
+# 3. Ensure ports 80 and 443 are open (Caddy handles SSL automatically)
+# 4. The frontend runs on port 3080 by default (docker-compose)
+#
+# Common Issues:
+# - "Connection refused": Check if the frontend container is running
+# - "Bad gateway": Check if localhost:3080 is accessible
+# - "SSL error": Make sure ports 80/443 are open for Let's Encrypt
 
-# Option 1: With domain name
-ai-photo-edit.yourdomain.com {
-    reverse_proxy localhost:3080
+# ============================================
+# OPTION 1: Domain with automatic HTTPS (recommended)
+# ============================================
+# Replace with your actual domain
+your-subdomain.yourdomain.com {
+    # Reverse proxy to frontend (nginx serves both frontend and proxies API)
+    reverse_proxy localhost:3080 {
+        # Health checks
+        health_uri /health
+        health_interval 30s
+        health_timeout 10s
+
+        # Headers for proper proxying
+        header_up Host {upstream_hostport}
+        header_up X-Real-IP {remote_host}
+        header_up X-Forwarded-For {remote_host}
+        header_up X-Forwarded-Proto {scheme}
+    }
+
+    # Enable compression
+    encode gzip zstd
+
+    # Logging (optional - uncomment for debugging)
+    # log {
+    #     output file /var/log/caddy/access.log
+    #     format json
+    # }
 }
 
-# Option 2: IP:Port (comment out option 1 if using this)
+# ============================================
+# OPTION 2: IP address or localhost (no HTTPS)
+# ============================================
+# Uncomment this block and comment out Option 1 if you don't have a domain
+# or want to test locally
+
 # :8080 {
-#     reverse_proxy localhost:3080
+#     reverse_proxy localhost:3080 {
+#         header_up Host {upstream_hostport}
+#         header_up X-Real-IP {remote_host}
+#         header_up X-Forwarded-For {remote_host}
+#     }
+#     encode gzip zstd
 # }
 
-# Note: You don't need to specify /projects, /edits, etc.
-# The frontend's nginx is already configured to proxy those to the backend internally.
+# ============================================
+# OPTION 3: Multiple subdomains
+# ============================================
+# If you want both www and non-www versions
+
+# yourdomain.com, www.yourdomain.com {
+#     reverse_proxy localhost:3080 {
+#         header_up Host {upstream_hostport}
+#         header_up X-Real-IP {remote_host}
+#         header_up X-Forwarded-For {remote_host}
+#         header_up X-Forwarded-Proto {scheme}
+#     }
+#     encode gzip zstd
+# }
+
+# ============================================
+# OPTION 4: Behind another reverse proxy (Cloudflare, etc.)
+# ============================================
+# Use this if Caddy is behind Cloudflare or another proxy
+
+# your-subdomain.yourdomain.com {
+#     # Trust proxy headers from upstream
+#     servers {
+#         trusted_proxies static 173.245.48.0/20 103.21.244.0/22 103.22.200.0/22 103.31.4.0/22 141.101.64.0/18 108.162.192.0/18 190.93.240.0/20 188.114.96.0/20 197.234.240.0/22 198.41.128.0/17 162.158.0.0/15 104.16.0.0/13 104.24.0.0/14 172.64.0.0/13 131.0.72.0/22
+#     }
+#
+#     reverse_proxy localhost:3080 {
+#         header_up Host {upstream_hostport}
+#         header_up X-Real-IP {http.request.header.CF-Connecting-IP}
+#         header_up X-Forwarded-For {http.request.header.CF-Connecting-IP}
+#         header_up X-Forwarded-Proto {scheme}
+#     }
+#     encode gzip zstd
+# }
+
+# ============================================
+# TROUBLESHOOTING
+# ============================================
+#
+# 1. Check Caddy logs:
+#    docker logs caddy
+#    OR: journalctl -u caddy -f
+#
+# 2. Test backend connectivity:
+#    curl -I http://localhost:3080
+#
+# 3. Check DNS resolution:
+#    dig your-subdomain.yourdomain.com
+#    nslookup your-subdomain.yourdomain.com
+#
+# 4. Verify ports are open:
+#    sudo netstat -tlnp | grep -E ':(80|443|3080)'
+#
+# 5. Check firewall:
+#    sudo ufw status
+#    sudo iptables -L -n
+#
+# 6. For Let's Encrypt issues:
+#    - Ensure ports 80 and 443 are accessible from internet
+#    - Check if domain resolves to your server's IP
+#    - Try: caddy validate --config /path/to/Caddyfile
+#
+# 7. Force reload Caddy config:
+#    caddy reload --config /path/to/Caddyfile
diff --git a/backend/app/routers/tools.py b/backend/app/routers/tools.py
index 062e2bf..c6eed00 100644
--- a/backend/app/routers/tools.py
+++ b/backend/app/routers/tools.py
@@ -128,53 +128,171 @@ async def inpaint_base64(request: InpaintRequest):
         raise HTTPException(status_code=500, detail=str(e))
 
 
+class RemoveBackgroundRequestV2(BaseModel):
+    image: str  # Base64 encoded image
+    model: Optional[str] = "auto"  # "auto", "u2net", "rembg", "birefnet"
+
+
 @router.post("/remove-background-base64")
 async def remove_background_base64(request: RemoveBackgroundRequest):
     """
-    Remove background from a base64 encoded image using rembg with BiRefNet.
-    BiRefNet is state-of-the-art for background removal (better than u2net).
+    Remove background from a base64 encoded image.
+    Tries multiple methods: U2Net (direct), rembg with BiRefNet, rembg default.
     Returns base64 encoded PNG with transparent background.
     Used by miniPaint frontend.
     """
-    try:
-        from rembg import remove, new_session
-    except ImportError:
-        raise HTTPException(
-            status_code=500,
-            detail="rembg not installed. Run: pip install rembg"
-        )
-
     try:
         # Decode base64 image
         image_bytes = base64.b64decode(request.image)
+        img = Image.open(BytesIO(image_bytes)).convert('RGB')
 
-        # Use BiRefNet model for best quality (state-of-the-art)
-        # Falls back to default model if BiRefNet not available
+        result_bytes = None
+        method_used = None
+
+        # Try U2Net first (direct implementation, no rembg dependency issues)
         try:
-            session = new_session("birefnet-general")
-            result_bytes = remove(image_bytes, session=session)
-        except Exception:
-            # Fallback to default model
-            result_bytes = remove(image_bytes)
+            result_bytes = await _remove_background_u2net(img)
+            method_used = "u2net"
+        except Exception as e:
+            print(f"U2Net failed: {e}")
+
+        # Fall back to rembg if U2Net failed
+        if result_bytes is None:
+            try:
+                from rembg import remove, new_session
+                try:
+                    session = new_session("birefnet-general")
+                    result_bytes = remove(image_bytes, session=session)
+                    method_used = "birefnet"
+                except Exception:
+                    result_bytes = remove(image_bytes)
+                    method_used = "rembg-default"
+            except ImportError:
+                pass
+            except Exception as e:
+                print(f"rembg failed: {e}")
+
+        if result_bytes is None:
+            raise HTTPException(
+                status_code=500,
+                detail="No background removal method available. Install u2net or rembg."
+            )
 
         # Convert result to base64
         result_b64 = base64.b64encode(result_bytes).decode('utf-8')
 
         # Get dimensions
-        img = Image.open(BytesIO(result_bytes))
+        result_img = Image.open(BytesIO(result_bytes))
 
         return {
             "result": result_b64,
-            "width": img.width,
-            "height": img.height
+            "width": result_img.width,
+            "height": result_img.height,
+            "method": method_used
         }
 
+    except HTTPException:
+        raise
     except Exception as e:
         import traceback
         traceback.print_exc()
         raise HTTPException(status_code=500, detail=str(e))
 
 
+# Global U2Net model cache
+_u2net_model = None
+
+
+async def _remove_background_u2net(img: Image.Image) -> bytes:
+    """
+    Remove background using U2Net model directly.
+    This avoids rembg dependency issues while providing good quality.
+    """
+    global _u2net_model
+
+    import torch
+    from pathlib import Path
+
+    # Check for U2Net model
+    models_dir = Path('/app/data/models')
+    u2net_path = models_dir / 'u2net.pth'
+
+    # Also check alternative names
+    if not u2net_path.exists():
+        for alt_name in ['u2net.onnx', 'u2netp.pth', 'u2net_human_seg.pth']:
+            alt_path = models_dir / alt_name
+            if alt_path.exists():
+                u2net_path = alt_path
+                break
+
+    if not u2net_path.exists():
+        raise FileNotFoundError(
+            f"U2Net model not found at {u2net_path}. "
+            "Download from: https://github.com/xuebinqin/U-2-Net"
+        )
+
+    # Load model if not cached
+    if _u2net_model is None:
+        print(f"Loading U2Net model from {u2net_path}")
+
+        if str(u2net_path).endswith('.onnx'):
+            # Use ONNX runtime
+            import onnxruntime as ort
+            _u2net_model = ort.InferenceSession(str(u2net_path))
+        else:
+            # Use PyTorch
+            from app.services.u2net_model import U2NET
+            _u2net_model = U2NET(3, 1)
+            _u2net_model.load_state_dict(torch.load(str(u2net_path), map_location='cpu'))
+            _u2net_model.eval()
+
+        print("U2Net model loaded")
+
+    # Preprocess image
+    img_np = np.array(img)
+    original_size = img.size
+
+    # Resize to model input size
+    input_size = 320
+    img_resized = img.resize((input_size, input_size), Image.Resampling.BILINEAR)
+    img_np = np.array(img_resized).astype(np.float32)
+
+    # Normalize
+    img_np = img_np / 255.0
+    img_np = (img_np - [0.485, 0.456, 0.406]) / [0.229, 0.224, 0.225]
+    img_np = img_np.transpose(2, 0, 1)  # HWC to CHW
+    img_np = np.expand_dims(img_np, 0)  # Add batch dimension
+
+    # Run inference
+    if hasattr(_u2net_model, 'run'):
+        # ONNX runtime
+        input_name = _u2net_model.get_inputs()[0].name
+        outputs = _u2net_model.run(None, {input_name: img_np})
+        mask = outputs[0][0, 0]
+    else:
+        # PyTorch
+        with torch.no_grad():
+            input_tensor = torch.from_numpy(img_np).float()
+            d1, d2, d3, d4, d5, d6, d7 = _u2net_model(input_tensor)
+            mask = d1[0, 0].numpy()
+
+    # Post-process mask
+    mask = (mask - mask.min()) / (mask.max() - mask.min() + 1e-8)
+    mask = (mask * 255).astype(np.uint8)
+
+    # Resize mask back to original size
+    mask_img = Image.fromarray(mask).resize(original_size, Image.Resampling.BILINEAR)
+
+    # Apply mask to original image
+    result = img.convert('RGBA')
+    result.putalpha(mask_img)
+
+    # Save to bytes
+    buffer = BytesIO()
+    result.save(buffer, format='PNG')
+    return buffer.getvalue()
+
+
 @router.post("/remove-background")
 async def remove_background(
     project_id: Optional[int] = Form(None),
diff --git a/backend/app/services/u2net_model.py b/backend/app/services/u2net_model.py
new file mode 100644
index 0000000..fa269b1
--- /dev/null
+++ b/backend/app/services/u2net_model.py
@@ -0,0 +1,500 @@
+"""
+U2Net Model Definition for Background Removal
+Based on: https://github.com/xuebinqin/U-2-Net
+
+This is a simplified implementation that works with the standard U2Net weights.
+"""
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class REBNCONV(nn.Module):
+    def __init__(self, in_ch=3, out_ch=3, dirate=1):
+        super(REBNCONV, self).__init__()
+
+        self.conv_s1 = nn.Conv2d(in_ch, out_ch, 3, padding=1*dirate, dilation=1*dirate)
+        self.bn_s1 = nn.BatchNorm2d(out_ch)
+        self.relu_s1 = nn.ReLU(inplace=True)
+
+    def forward(self, x):
+        hx = x
+        xout = self.relu_s1(self.bn_s1(self.conv_s1(hx)))
+        return xout
+
+
+def _upsample_like(src, tar):
+    src = F.interpolate(src, size=tar.shape[2:], mode='bilinear', align_corners=False)
+    return src
+
+
+class RSU7(nn.Module):
+    def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
+        super(RSU7, self).__init__()
+
+        self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
+
+        self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
+        self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool5 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=1)
+
+        self.rebnconv7 = REBNCONV(mid_ch, mid_ch, dirate=2)
+
+        self.rebnconv6d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv5d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
+
+    def forward(self, x):
+        hx = x
+        hxin = self.rebnconvin(hx)
+
+        hx1 = self.rebnconv1(hxin)
+        hx = self.pool1(hx1)
+
+        hx2 = self.rebnconv2(hx)
+        hx = self.pool2(hx2)
+
+        hx3 = self.rebnconv3(hx)
+        hx = self.pool3(hx3)
+
+        hx4 = self.rebnconv4(hx)
+        hx = self.pool4(hx4)
+
+        hx5 = self.rebnconv5(hx)
+        hx = self.pool5(hx5)
+
+        hx6 = self.rebnconv6(hx)
+
+        hx7 = self.rebnconv7(hx6)
+
+        hx6d = self.rebnconv6d(torch.cat((hx7, hx6), 1))
+        hx6dup = _upsample_like(hx6d, hx5)
+
+        hx5d = self.rebnconv5d(torch.cat((hx6dup, hx5), 1))
+        hx5dup = _upsample_like(hx5d, hx4)
+
+        hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1))
+        hx4dup = _upsample_like(hx4d, hx3)
+
+        hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
+
+        return hx1d + hxin
+
+
+class RSU6(nn.Module):
+    def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
+        super(RSU6, self).__init__()
+
+        self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
+
+        self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
+        self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1)
+
+        self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=2)
+
+        self.rebnconv5d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
+
+    def forward(self, x):
+        hx = x
+        hxin = self.rebnconvin(hx)
+
+        hx1 = self.rebnconv1(hxin)
+        hx = self.pool1(hx1)
+
+        hx2 = self.rebnconv2(hx)
+        hx = self.pool2(hx2)
+
+        hx3 = self.rebnconv3(hx)
+        hx = self.pool3(hx3)
+
+        hx4 = self.rebnconv4(hx)
+        hx = self.pool4(hx4)
+
+        hx5 = self.rebnconv5(hx)
+
+        hx6 = self.rebnconv6(hx5)
+
+        hx5d = self.rebnconv5d(torch.cat((hx6, hx5), 1))
+        hx5dup = _upsample_like(hx5d, hx4)
+
+        hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1))
+        hx4dup = _upsample_like(hx4d, hx3)
+
+        hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
+
+        return hx1d + hxin
+
+
+class RSU5(nn.Module):
+    def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
+        super(RSU5, self).__init__()
+
+        self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
+
+        self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
+        self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
+
+        self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=2)
+
+        self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
+
+    def forward(self, x):
+        hx = x
+        hxin = self.rebnconvin(hx)
+
+        hx1 = self.rebnconv1(hxin)
+        hx = self.pool1(hx1)
+
+        hx2 = self.rebnconv2(hx)
+        hx = self.pool2(hx2)
+
+        hx3 = self.rebnconv3(hx)
+        hx = self.pool3(hx3)
+
+        hx4 = self.rebnconv4(hx)
+
+        hx5 = self.rebnconv5(hx4)
+
+        hx4d = self.rebnconv4d(torch.cat((hx5, hx4), 1))
+        hx4dup = _upsample_like(hx4d, hx3)
+
+        hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
+
+        return hx1d + hxin
+
+
+class RSU4(nn.Module):
+    def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
+        super(RSU4, self).__init__()
+
+        self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
+
+        self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
+        self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
+        self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
+
+        self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=2)
+
+        self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
+        self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
+
+    def forward(self, x):
+        hx = x
+        hxin = self.rebnconvin(hx)
+
+        hx1 = self.rebnconv1(hxin)
+        hx = self.pool1(hx1)
+
+        hx2 = self.rebnconv2(hx)
+        hx = self.pool2(hx2)
+
+        hx3 = self.rebnconv3(hx)
+
+        hx4 = self.rebnconv4(hx3)
+
+        hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
+
+        return hx1d + hxin
+
+
+class RSU4F(nn.Module):
+    def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
+        super(RSU4F, self).__init__()
+
+        self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
+
+        self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
+        self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=2)
+        self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=4)
+
+        self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=8)
+
+        self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=4)
+        self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=2)
+        self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
+
+    def forward(self, x):
+        hx = x
+        hxin = self.rebnconvin(hx)
+
+        hx1 = self.rebnconv1(hxin)
+        hx2 = self.rebnconv2(hx1)
+        hx3 = self.rebnconv3(hx2)
+
+        hx4 = self.rebnconv4(hx3)
+
+        hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1))
+        hx2d = self.rebnconv2d(torch.cat((hx3d, hx2), 1))
+        hx1d = self.rebnconv1d(torch.cat((hx2d, hx1), 1))
+
+        return hx1d + hxin
+
+
+class U2NET(nn.Module):
+    def __init__(self, in_ch=3, out_ch=1):
+        super(U2NET, self).__init__()
+
+        self.stage1 = RSU7(in_ch, 32, 64)
+        self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage2 = RSU6(64, 32, 128)
+        self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage3 = RSU5(128, 64, 256)
+        self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage4 = RSU4(256, 128, 512)
+        self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage5 = RSU4F(512, 256, 512)
+        self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage6 = RSU4F(512, 256, 512)
+
+        # decoder
+        self.stage5d = RSU4F(1024, 256, 512)
+        self.stage4d = RSU4(1024, 128, 256)
+        self.stage3d = RSU5(512, 64, 128)
+        self.stage2d = RSU6(256, 32, 64)
+        self.stage1d = RSU7(128, 16, 64)
+
+        self.side1 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side2 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side3 = nn.Conv2d(128, out_ch, 3, padding=1)
+        self.side4 = nn.Conv2d(256, out_ch, 3, padding=1)
+        self.side5 = nn.Conv2d(512, out_ch, 3, padding=1)
+        self.side6 = nn.Conv2d(512, out_ch, 3, padding=1)
+
+        self.outconv = nn.Conv2d(6*out_ch, out_ch, 1)
+
+    def forward(self, x):
+        hx = x
+
+        # stage 1
+        hx1 = self.stage1(hx)
+        hx = self.pool12(hx1)
+
+        # stage 2
+        hx2 = self.stage2(hx)
+        hx = self.pool23(hx2)
+
+        # stage 3
+        hx3 = self.stage3(hx)
+        hx = self.pool34(hx3)
+
+        # stage 4
+        hx4 = self.stage4(hx)
+        hx = self.pool45(hx4)
+
+        # stage 5
+        hx5 = self.stage5(hx)
+        hx = self.pool56(hx5)
+
+        # stage 6
+        hx6 = self.stage6(hx)
+        hx6up = _upsample_like(hx6, hx5)
+
+        # decoder
+        hx5d = self.stage5d(torch.cat((hx6up, hx5), 1))
+        hx5dup = _upsample_like(hx5d, hx4)
+
+        hx4d = self.stage4d(torch.cat((hx5dup, hx4), 1))
+        hx4dup = _upsample_like(hx4d, hx3)
+
+        hx3d = self.stage3d(torch.cat((hx4dup, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.stage2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.stage1d(torch.cat((hx2dup, hx1), 1))
+
+        # side output
+        d1 = self.side1(hx1d)
+
+        d2 = self.side2(hx2d)
+        d2 = _upsample_like(d2, d1)
+
+        d3 = self.side3(hx3d)
+        d3 = _upsample_like(d3, d1)
+
+        d4 = self.side4(hx4d)
+        d4 = _upsample_like(d4, d1)
+
+        d5 = self.side5(hx5d)
+        d5 = _upsample_like(d5, d1)
+
+        d6 = self.side6(hx6)
+        d6 = _upsample_like(d6, d1)
+
+        d0 = self.outconv(torch.cat((d1, d2, d3, d4, d5, d6), 1))
+
+        return torch.sigmoid(d0), torch.sigmoid(d1), torch.sigmoid(d2), torch.sigmoid(d3), torch.sigmoid(d4), torch.sigmoid(d5), torch.sigmoid(d6)
+
+
+class U2NETP(nn.Module):
+    """Smaller/faster U2Net variant (u2netp)"""
+
+    def __init__(self, in_ch=3, out_ch=1):
+        super(U2NETP, self).__init__()
+
+        self.stage1 = RSU7(in_ch, 16, 64)
+        self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage2 = RSU6(64, 16, 64)
+        self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage3 = RSU5(64, 16, 64)
+        self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage4 = RSU4(64, 16, 64)
+        self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage5 = RSU4F(64, 16, 64)
+        self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
+
+        self.stage6 = RSU4F(64, 16, 64)
+
+        # decoder
+        self.stage5d = RSU4F(128, 16, 64)
+        self.stage4d = RSU4(128, 16, 64)
+        self.stage3d = RSU5(128, 16, 64)
+        self.stage2d = RSU6(128, 16, 64)
+        self.stage1d = RSU7(128, 16, 64)
+
+        self.side1 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side2 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side3 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side4 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side5 = nn.Conv2d(64, out_ch, 3, padding=1)
+        self.side6 = nn.Conv2d(64, out_ch, 3, padding=1)
+
+        self.outconv = nn.Conv2d(6*out_ch, out_ch, 1)
+
+    def forward(self, x):
+        hx = x
+
+        hx1 = self.stage1(hx)
+        hx = self.pool12(hx1)
+
+        hx2 = self.stage2(hx)
+        hx = self.pool23(hx2)
+
+        hx3 = self.stage3(hx)
+        hx = self.pool34(hx3)
+
+        hx4 = self.stage4(hx)
+        hx = self.pool45(hx4)
+
+        hx5 = self.stage5(hx)
+        hx = self.pool56(hx5)
+
+        hx6 = self.stage6(hx)
+        hx6up = _upsample_like(hx6, hx5)
+
+        hx5d = self.stage5d(torch.cat((hx6up, hx5), 1))
+        hx5dup = _upsample_like(hx5d, hx4)
+
+        hx4d = self.stage4d(torch.cat((hx5dup, hx4), 1))
+        hx4dup = _upsample_like(hx4d, hx3)
+
+        hx3d = self.stage3d(torch.cat((hx4dup, hx3), 1))
+        hx3dup = _upsample_like(hx3d, hx2)
+
+        hx2d = self.stage2d(torch.cat((hx3dup, hx2), 1))
+        hx2dup = _upsample_like(hx2d, hx1)
+
+        hx1d = self.stage1d(torch.cat((hx2dup, hx1), 1))
+
+        d1 = self.side1(hx1d)
+
+        d2 = self.side2(hx2d)
+        d2 = _upsample_like(d2, d1)
+
+        d3 = self.side3(hx3d)
+        d3 = _upsample_like(d3, d1)
+
+        d4 = self.side4(hx4d)
+        d4 = _upsample_like(d4, d1)
+
+        d5 = self.side5(hx5d)
+        d5 = _upsample_like(d5, d1)
+
+        d6 = self.side6(hx6)
+        d6 = _upsample_like(d6, d1)
+
+        d0 = self.outconv(torch.cat((d1, d2, d3, d4, d5, d6), 1))
+
+        return torch.sigmoid(d0), torch.sigmoid(d1), torch.sigmoid(d2), torch.sigmoid(d3), torch.sigmoid(d4), torch.sigmoid(d5), torch.sigmoid(d6)
diff --git a/frontend/src/css/popup.css b/frontend/src/css/popup.css
index 2ac6e82..5a72e74 100644
--- a/frontend/src/css/popup.css
+++ b/frontend/src/css/popup.css
@@ -245,6 +245,143 @@
 	margin-right: auto;
 }
 
+/* Shape/Library Tabs */
+#popups .popup .shape-tabs {
+	display: flex;
+	gap: 0;
+	margin-bottom: 1rem;
+	border-bottom: 2px solid var(--border-color);
+}
+
+#popups .popup .shape-tab {
+	padding: 0.8rem 1.5rem;
+	background: transparent;
+	border: none;
+	border-bottom: 2px solid transparent;
+	margin-bottom: -2px;
+	cursor: pointer;
+	color: var(--text-color-muted);
+	font-size: 1rem;
+	transition: color 0.2s, border-color 0.2s;
+}
+
+#popups .popup .shape-tab:hover {
+	color: var(--text-color);
+}
+
+#popups .popup .shape-tab.active {
+	color: var(--link-color);
+	border-bottom-color: var(--link-color);
+}
+
+#popups .popup .library-loading {
+	text-align: center;
+	padding: 2rem;
+	color: var(--text-color-muted);
+}
+
+/* My Library Browser Styles */
+#popups .popup .library-browser {
+	max-height: calc(60vh - 100px);
+	overflow-y: auto;
+}
+
+#popups .popup .library-categories {
+	display: flex;
+	flex-direction: column;
+	gap: 1.5rem;
+}
+
+#popups .popup .library-category h3 {
+	color: var(--text-color);
+	font-size: 1.4rem;
+	margin-bottom: 0.8rem;
+	padding-bottom: 0.4rem;
+	border-bottom: 1px solid var(--border-color);
+}
+
+#popups .popup .library-items {
+	display: flex;
+	flex-wrap: wrap;
+	gap: 1rem;
+}
+
+#popups .popup .library-item {
+	display: flex;
+	flex-direction: column;
+	align-items: center;
+	width: 120px;
+	padding: 0.8rem;
+	background: var(--input-background-color);
+	border: 1px solid var(--border-color);
+	border-radius: 4px;
+	cursor: pointer;
+	transition: background 0.2s, border-color 0.2s;
+}
+
+#popups .popup .library-item:hover {
+	background: var(--input-background-color-hover);
+	border-color: var(--link-color);
+}
+
+#popups .popup .library-item img {
+	width: 100px;
+	height: 80px;
+	object-fit: contain;
+	background: repeating-conic-gradient(#666 0% 25%, #888 0% 50%) 50% / 10px 10px;
+	border-radius: 2px;
+	margin-bottom: 0.5rem;
+}
+
+#popups .popup .library-item-name {
+	font-size: 0.85rem;
+	text-align: center;
+	color: var(--text-color);
+	white-space: nowrap;
+	overflow: hidden;
+	text-overflow: ellipsis;
+	width: 100%;
+	margin-bottom: 0.5rem;
+}
+
+#popups .popup .library-item-actions {
+	display: flex;
+	gap: 0.5rem;
+}
+
+#popups .popup .library-item-actions button {
+	font-size: 0.75rem;
+	padding: 0.3rem 0.6rem;
+	background: var(--button-background-color);
+	border: 1px solid var(--border-color);
+	border-radius: 3px;
+	cursor: pointer;
+	color: var(--text-color);
+}
+
+#popups .popup .library-item-actions .insert-btn {
+	background: #2a6d2a;
+}
+
+#popups .popup .library-item-actions .insert-btn:hover {
+	background: #3a8d3a;
+}
+
+#popups .popup .library-item-actions .delete-btn {
+	background: #6d2a2a;
+}
+
+#popups .popup .library-item-actions .delete-btn:hover {
+	background: #8d3a3a;
+}
+
+/* Library empty state */
+#popups .popup .library-empty {
+	text-align: center;
+	padding: 2rem;
+	color: var(--text-color-muted);
+}
+
 @media screen and (max-width:500px){
 	#popups .popup {
 		max-height: calc(80vh - 20px); /* mobile phones has bottom menu */
diff --git a/frontend/src/js/modules/file/my_library.js b/frontend/src/js/modules/file/my_library.js
index 4d60869..40d9574 100644
--- a/frontend/src/js/modules/file/my_library.js
+++ b/frontend/src/js/modules/file/my_library.js
@@ -334,7 +334,7 @@ class File_my_library_class {
             title: 'My Library (' + assets.length + ' assets)',
             params: [],
             html: html,
-            className: 'library-dialog',
+            className: 'wide',
             on_load: function(el) {
                 // Add click handlers
                 el.querySelectorAll('.insert-btn').forEach(function(btn) {
diff --git a/frontend/src/js/tools/ai_inpaint.js b/frontend/src/js/tools/ai_inpaint.js
index 1bff763..54e2f56 100644
--- a/frontend/src/js/tools/ai_inpaint.js
+++ b/frontend/src/js/tools/ai_inpaint.js
@@ -49,14 +49,20 @@ class Ai_inpaint_class extends Base_tools_class {
         }
 
         var settings = {
-            title: 'AI Inpaint',
+            title: 'AI Edit Selection',
             params: [
+                {
+                    name: "mode",
+                    title: "Edit Mode:",
+                    value: "inpaint",
+                    values: ["inpaint", "transform"]
+                },
                 {
                     name: "prompt",
-                    title: "Describe what you want:",
+                    title: "AI Inpaint - Describe replacement:",
                     type: "textarea",
                     value: "",
-                    placeholder: "e.g., 'a red rose', 'remove the object', 'blue sky with clouds'"
+                    placeholder: "AI will REPLACE the selection with what you describe.\nExamples: 'a red rose', 'empty background', 'blue sky'"
                 },
                 {
                     name: "negative_prompt",
@@ -66,21 +72,164 @@ class Ai_inpaint_class extends Base_tools_class {
                 },
                 {
                     name: "strength",
-                    title: "Edit Strength:",
+                    title: "AI Edit Strength:",
                     type: "range",
                     value: 80,
                     range: [1, 100],
                     step: 1
+                },
+                {
+                    name: "scale",
+                    title: "Transform - Scale %:",
+                    type: "range",
+                    value: 100,
+                    range: [10, 200],
+                    step: 5
                 }
             ],
+            on_load: function(el) {
+                // Add info text
+                var infoDiv = document.createElement('div');
+                infoDiv.className = 'ai-inpaint-info';
+                infoDiv.innerHTML = '<p style="font-size:12px;color:#aaa;margin-bottom:10px;">' +
+                    '<strong>Inpaint Mode:</strong> AI replaces the selected area with generated content.<br>' +
+                    '<strong>Transform Mode:</strong> Scale, shrink, or enlarge the selection without AI.<br>' +
+                    '<em>Tip: To shrink something by 35%, use Transform mode with Scale at 65%.</em></p>';
+
+                var dialogContent = el.querySelector('.dialog_content');
+                if (dialogContent && dialogContent.firstChild) {
+                    dialogContent.insertBefore(infoDiv, dialogContent.firstChild);
+                }
+            },
             on_finish: async function (params) {
-                await _this.executeInpaint(params);
+                if (params.mode === 'transform') {
+                    await _this.executeTransform(params);
+                } else {
+                    await _this.executeInpaint(params);
+                }
             },
         };
 
         this.POP.show(settings);
     }
 
+    /**
+     * Execute transform operation (scale without AI)
+     */
+    async executeTransform(params) {
+        if (this.isProcessing) {
+            alertify.warning('Already processing... please wait');
+            return;
+        }
+
+        // Check if we have an image layer
+        if (config.layer.type != 'image') {
+            alertify.error('Please select an image layer');
+            return;
+        }
+
+        var maskCanvas = window.smartSelectMask?.canvas;
+        if (!maskCanvas) {
+            alertify.error('No selection mask found');
+            return;
+        }
+
+        this.isProcessing = true;
+        alertify.message('Transforming selection...');
+
+        try {
+            var layer = config.layer;
+            var scale = params.scale / 100;
+
+            // Get mask bounds
+            var maskCtx = maskCanvas.getContext('2d');
+            var imageData = maskCtx.getImageData(0, 0, maskCanvas.width, maskCanvas.height);
+            var minX = maskCanvas.width, minY = maskCanvas.height;
+            var maxX = 0, maxY = 0;
+
+            for (var y = 0; y < maskCanvas.height; y++) {
+                for (var x = 0; x < maskCanvas.width; x++) {
+                    var i = (y * maskCanvas.width + x) * 4;
+                    if (imageData.data[i] > 128) {
+                        minX = Math.min(minX, x);
+                        minY = Math.min(minY, y);
+                        maxX = Math.max(maxX, x);
+                        maxY = Math.max(maxY, y);
+                    }
+                }
+            }
+
+            if (maxX <= minX || maxY <= minY) {
+                throw new Error('Selection is too small');
+            }
+
+            var selWidth = maxX - minX + 1;
+            var selHeight = maxY - minY + 1;
+            var centerX = minX + selWidth / 2;
+            var centerY = minY + selHeight / 2;
+
+            // Extract selected pixels
+            var extractCanvas = document.createElement('canvas');
+            extractCanvas.width = layer.width_original;
+            extractCanvas.height = layer.height_original;
+            var extractCtx = extractCanvas.getContext('2d');
+            extractCtx.drawImage(layer.link, 0, 0);
+            extractCtx.globalCompositeOperation = 'destination-in';
+            extractCtx.drawImage(maskCanvas, 0, 0);
+
+            // Create result canvas
+            var resultCanvas = document.createElement('canvas');
+            resultCanvas.width = layer.width_original;
+            resultCanvas.height = layer.height_original;
+            var resultCtx = resultCanvas.getContext('2d');
+
+            // Draw original image
+            resultCtx.drawImage(layer.link, 0, 0);
+
+            // Remove original selection (create hole)
+            resultCtx.globalCompositeOperation = 'destination-out';
+            resultCtx.drawImage(maskCanvas, 0, 0);
+
+            // Calculate scaled dimensions
+            var newWidth = selWidth * scale;
+            var newHeight = selHeight * scale;
+            var newX = centerX - newWidth / 2;
+            var newY = centerY - newHeight / 2;
+
+            // Draw scaled selection back
+            resultCtx.globalCompositeOperation = 'source-over';
+
+            // Create temp canvas for just the selection
+            var selCanvas = document.createElement('canvas');
+            selCanvas.width = selWidth;
+            selCanvas.height = selHeight;
+            var selCtx = selCanvas.getContext('2d');
+            selCtx.drawImage(extractCanvas, minX, minY, selWidth, selHeight, 0, 0, selWidth, selHeight);
+
+            // Draw scaled
+            resultCtx.drawImage(selCanvas, 0, 0, selWidth, selHeight, newX, newY, newWidth, newHeight);
+
+            // Apply result
+            app.State.do_action(
+                new app.Actions.Bundle_action('transform_selection', 'Transform Selection', [
+                    new app.Actions.Update_layer_image_action(resultCanvas)
+                ])
+            );
+
+            // Clear selection
+            window.smartSelectMask = null;
+            config.need_render = true;
+
+            alertify.success('Transform complete! Selection scaled to ' + params.scale + '%');
+
+        } catch (error) {
+            console.error('Transform error:', error);
+            alertify.error('Transform failed: ' + error.message);
+        } finally {
+            this.isProcessing = false;
+        }
+    }
+
     /**
      * Execute the inpainting operation
      */
diff --git a/frontend/src/js/tools/shape.js b/frontend/src/js/tools/shape.js
index 2269df3..e48f792 100644
--- a/frontend/src/js/tools/shape.js
+++ b/frontend/src/js/tools/shape.js
@@ -4,6 +4,7 @@ import Base_tools_class from './../core/base-tools.js';
 import Base_layers_class from './../core/base-layers.js';
 import Dialog_class from './../libs/popup.js';
 import GUI_tools_class from './../core/gui/gui-tools.js';
+import File_my_library_class from './../modules/file/my_library.js';
 
 var instance = null;
 
@@ -21,11 +22,13 @@ class Shape_class extends Base_tools_class {
 		this.Base_layers = new Base_layers_class();
 		this.GUI_tools = new GUI_tools_class();
 		this.POP = new Dialog_class();
+		this.My_library = new File_my_library_class();
 		this.ctx = ctx;
 		this.name = 'shape';
 		this.layer = {};
 		this.preview_width = 150;
 		this.preview_height = 120;
+		this.activeTab = 'shapes'; // 'shapes' or 'library'
 
 		this.set_events();
 	}
@@ -53,35 +56,84 @@ class Shape_class extends Base_tools_class {
 
 	async show_shapes(){
 		var _this = this;
-		var html = '';
 
+		// Build tabs HTML
+		var tabsHtml = '<div class="shape-tabs">';
+		tabsHtml += '<button class="shape-tab active" data-tab="shapes">Built-in Shapes</button>';
+		tabsHtml += '<button class="shape-tab" data-tab="library">My Library</button>';
+		tabsHtml += '</div>';
+
+		// Build shapes HTML
+		var shapesHtml = '<div class="tab-content shapes-content">';
 		var data = this.get_shapes();
 
 		for (var i in data) {
-			html += '<div class="item">';
-			html += '	<canvas id="c_' + data[i].key + '" width="' + this.preview_width + '" height="'
+			shapesHtml += '<div class="item">';
+			shapesHtml += '	<canvas id="c_' + data[i].key + '" width="' + this.preview_width + '" height="'
 				+ this.preview_height + '" class="effectsPreview" data-key="'
 				+ data[i].key + '"></canvas>';
-			html += '<div class="preview-item-title">' + data[i].title + '</div>';
-			html += '</div>';
+			shapesHtml += '<div class="preview-item-title">' + data[i].title + '</div>';
+			shapesHtml += '</div>';
 		}
 		for (var i = 0; i < 4; i++) {
-			html += '<div class="item"></div>';
+			shapesHtml += '<div class="item"></div>';
 		}
+		shapesHtml += '</div>';
+
+		// Build library HTML placeholder
+		var libraryHtml = '<div class="tab-content library-content" style="display:none;">';
+		libraryHtml += '<div class="library-loading">Loading library...</div>';
+		libraryHtml += '</div>';
 
 		var settings = {
-			title: 'Shapes',
+			title: 'Shapes & Library',
 			className: 'wide',
 			on_load: function (params, popup) {
-				var node = document.createElement("div");
-				node.classList.add('flex-container');
-				node.innerHTML = html;
-				popup.el.querySelector('.dialog_content').appendChild(node);
-				//events
+				// Add tabs
+				var tabsNode = document.createElement("div");
+				tabsNode.innerHTML = tabsHtml;
+				popup.el.querySelector('.dialog_content').insertBefore(tabsNode, popup.el.querySelector('.dialog_content').firstChild);
+
+				// Add shapes container
+				var shapesNode = document.createElement("div");
+				shapesNode.classList.add('flex-container');
+				shapesNode.innerHTML = shapesHtml;
+				popup.el.querySelector('.dialog_content').appendChild(shapesNode);
+
+				// Add library container
+				var libraryNode = document.createElement("div");
+				libraryNode.innerHTML = libraryHtml;
+				popup.el.querySelector('.dialog_content').appendChild(libraryNode);
+
+				// Tab click events
+				var tabs = popup.el.querySelectorAll('.shape-tab');
+				tabs.forEach(function(tab) {
+					tab.addEventListener('click', function() {
+						var targetTab = this.dataset.tab;
+
+						// Update active tab
+						tabs.forEach(t => t.classList.remove('active'));
+						this.classList.add('active');
+
+						// Show/hide content
+						var shapesContent = popup.el.querySelector('.shapes-content');
+						var libraryContent = popup.el.querySelector('.library-content');
+
+						if (targetTab === 'shapes') {
+							shapesContent.style.display = '';
+							libraryContent.style.display = 'none';
+						} else {
+							shapesContent.style.display = 'none';
+							libraryContent.style.display = '';
+							_this.loadLibraryContent(libraryContent);
+						}
+					});
+				});
+
+				// Shape click events
 				var targets = popup.el.querySelectorAll('.item canvas');
 				for (var i = 0; i < targets.length; i++) {
 					targets[i].addEventListener('click', function (event) {
-						//we have click
 						_this.GUI_tools.activate_tool(this.dataset.key);
 						_this.POP.hide();
 					});
@@ -106,6 +158,89 @@ class Shape_class extends Base_tools_class {
 		}
 	}
 
+	/**
+	 * Load library content into the library tab
+	 */
+	loadLibraryContent(container) {
+		var _this = this;
+
+		this.My_library.getAllAssets(function(assets) {
+			var html = '';
+
+			if (assets.length === 0) {
+				html = '<div class="library-empty">';
+				html += '<p>Your library is empty.</p>';
+				html += '<p>Use <strong>File > My Library > Save to Library</strong> to add assets.</p>';
+				html += '</div>';
+			} else {
+				// Group by category
+				var categories = {};
+				assets.forEach(function(asset) {
+					var cat = asset.category || 'General';
+					if (!categories[cat]) categories[cat] = [];
+					categories[cat].push(asset);
+				});
+
+				html = '<div class="library-browser">';
+				html += '<div class="library-categories">';
+
+				for (var cat in categories) {
+					html += '<div class="library-category">';
+					html += '<h3>' + cat + ' (' + categories[cat].length + ')</h3>';
+					html += '<div class="library-items">';
+
+					categories[cat].forEach(function(asset) {
+						html += '<div class="library-item" data-id="' + asset.id + '">';
+						html += '<img src="' + asset.thumbnail + '" alt="' + asset.name + '" title="' + asset.name + '">';
+						html += '<div class="library-item-name">' + asset.name + '</div>';
+						html += '<div class="library-item-actions">';
+						html += '<button class="insert-btn" data-id="' + asset.id + '">Insert</button>';
+						html += '<button class="delete-btn" data-id="' + asset.id + '">Delete</button>';
+						html += '</div>';
+						html += '</div>';
+					});
+
+					html += '</div></div>';
+				}
+
+				html += '</div></div>';
+			}
+
+			container.innerHTML = html;
+
+			// Add event handlers for library items
+			container.querySelectorAll('.insert-btn').forEach(function(btn) {
+				btn.addEventListener('click', function(e) {
+					e.stopPropagation();
+					var id = parseInt(this.dataset.id);
+					_this.My_library.insertAsset(id);
+					_this.POP.hide();
+				});
+			});
+
+			container.querySelectorAll('.delete-btn').forEach(function(btn) {
+				btn.addEventListener('click', function(e) {
+					e.stopPropagation();
+					var id = parseInt(this.dataset.id);
+					if (confirm('Delete this asset?')) {
+						_this.My_library.deleteAsset(id, function() {
+							_this.loadLibraryContent(container);
+						});
+					}
+				});
+			});
+
+			// Double-click to insert
+			container.querySelectorAll('.library-item').forEach(function(item) {
+				item.addEventListener('dblclick', function() {
+					var id = parseInt(this.dataset.id);
+					_this.My_library.insertAsset(id);
+					_this.POP.hide();
+				});
+			});
+		});
+	}
+
 	render(ctx, layer) {
 
 	}
diff --git a/scripts/download_u2net_model.py b/scripts/download_u2net_model.py
new file mode 100644
index 0000000..3f44b75
--- /dev/null
+++ b/scripts/download_u2net_model.py
@@ -0,0 +1,133 @@
+#!/usr/bin/env python3
+"""
+Download U2Net model for background removal.
+
+U2Net is a deep learning model for salient object detection,
+commonly used for background removal tasks.
+
+Usage:
+    python download_u2net_model.py [model_type]
+
+Model types:
+    u2net      - Full U2Net model (~176MB, best quality)
+    u2netp     - Lightweight U2Net (~4MB, faster, good quality)
+    u2net_human_seg - Optimized for human segmentation (~176MB)
+
+Default: u2netp (good balance of quality and speed)
+"""
+
+import os
+import sys
+import urllib.request
+from pathlib import Path
+
+# Model URLs (from official U2Net repository releases)
+MODEL_URLS = {
+    'u2net': {
+        'url': 'https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net.onnx',
+        'filename': 'u2net.onnx',
+        'size_mb': 176
+    },
+    'u2netp': {
+        'url': 'https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2netp.onnx',
+        'filename': 'u2netp.onnx',
+        'size_mb': 4
+    },
+    'u2net_human_seg': {
+        'url': 'https://github.com/danielgatis/rembg/releases/download/v0.0.0/u2net_human_seg.onnx',
+        'filename': 'u2net_human_seg.onnx',
+        'size_mb': 176
+    }
+}
+
+
+def download_with_progress(url: str, dest_path: Path, expected_size_mb: int):
+    """Download file with progress indicator."""
+
+    print(f"Downloading from: {url}")
+    print(f"Expected size: ~{expected_size_mb}MB")
+
+    def progress_hook(count, block_size, total_size):
+        if total_size > 0:
+            percent = min(100, count * block_size * 100 // total_size)
+            downloaded_mb = count * block_size / (1024 * 1024)
+            total_mb = total_size / (1024 * 1024)
+            sys.stdout.write(f"\rProgress: {percent}% ({downloaded_mb:.1f}/{total_mb:.1f} MB)")
+            sys.stdout.flush()
+
+    try:
+        urllib.request.urlretrieve(url, str(dest_path), progress_hook)
+        print("\nDownload complete!")
+        return True
+    except Exception as e:
+        print(f"\nDownload failed: {e}")
+        return False
+
+
+def main():
+    # Determine model type
+    model_type = 'u2netp'  # Default to lightweight model
+    if len(sys.argv) > 1:
+        model_type = sys.argv[1].lower()
+
+    if model_type not in MODEL_URLS:
+        print(f"Unknown model type: {model_type}")
+        print(f"Available models: {', '.join(MODEL_URLS.keys())}")
+        sys.exit(1)
+
+    model_info = MODEL_URLS[model_type]
+
+    # Determine models directory
+    # Check if running in Docker container
+    if os.path.exists('/app/data/models'):
+        models_dir = Path('/app/data/models')
+    else:
+        # Local development
+        script_dir = Path(__file__).parent
+        models_dir = script_dir.parent / 'data' / 'models'
+
+    models_dir.mkdir(parents=True, exist_ok=True)
+
+    dest_path = models_dir / model_info['filename']
+
+    # Check if already downloaded
+    if dest_path.exists():
+        print(f"Model already exists at: {dest_path}")
+        print("Delete the file to re-download.")
+        return
+
+    print(f"Downloading U2Net model: {model_type}")
+    print(f"Destination: {dest_path}")
+    print("")
+
+    success = download_with_progress(
+        model_info['url'],
+        dest_path,
+        model_info['size_mb']
+    )
+
+    if success:
+        # Create symlink for easier access
+        symlink_path = models_dir / 'u2net.onnx'
+        if not symlink_path.exists() or symlink_path.is_symlink():
+            if symlink_path.is_symlink():
+                symlink_path.unlink()
+            try:
+                symlink_path.symlink_to(dest_path.name)
+                print(f"Created symlink: {symlink_path} -> {dest_path.name}")
+            except OSError:
+                # Symlinks may not work on all systems
+                pass
+
+        print(f"\nU2Net model ({model_type}) downloaded successfully!")
+        print(f"Location: {dest_path}")
+        print("\nYou can now use background removal in the application.")
+    else:
+        print("\nFailed to download model. Please try again or download manually from:")
+        print(f"  {model_info['url']}")
+        print(f"  Save to: {dest_path}")
+        sys.exit(1)
+
+
+if __name__ == '__main__':
+    main()