diff --git a/src/lib/images.ts b/src/lib/images.ts
index fdacfca..93a9a0a 100644
--- a/src/lib/images.ts
+++ b/src/lib/images.ts
@@ -1,40 +1,29 @@
-// import * as tf from "@tensorflow/tfjs-node";
-// import * as nsfwjs from "nsfwjs";
 import sharp from "sharp";
 import { fileTypeFromBuffer } from "file-type";
 
 const MIN_IMAGE_DIMENSIONS = 128;
 const MAX_IMAGE_DIMENSIONS = 1024;
 const MAX_IMAGE_SIZE = 1024 * 1024; // 1 MB
-
-// const THRESHOLD = 0.5;
-
-// tf.enableProdMode();
-
-// Load NSFW.JS model
-// let _model: nsfwjs.NSFWJS | undefined = undefined;
-
-// async function loadModel() {
-// 	if (!_model) {
-// 		const model = await nsfwjs.load("MobileNetV2Mid");
-// 		_model = model;
-// 	}
-// 	return _model!;
-// }
+const ALLOWED_MIME_TYPES = ["image/jpeg", "image/png", "image/gif", "image/webp"];
 
 export async function validateImage(file: File): Promise<{ valid: boolean; error?: string; status?: number }> {
 	if (!file || file.size == 0) return { valid: false, error: "Empty image file" };
-	if (file.size > MAX_IMAGE_SIZE)
-		return { valid: false, error: `One or more of your images are too large. Maximum size is ${MAX_IMAGE_SIZE / (1024 * 1024)}MB` };
+	if (file.size > MAX_IMAGE_SIZE) return { valid: false, error: `Image too large. Maximum size is ${MAX_IMAGE_SIZE / (1024 * 1024)}MB` };
 
 	try {
 		const buffer = Buffer.from(await file.arrayBuffer());
 
 		// Check mime type
 		const fileType = await fileTypeFromBuffer(buffer);
-		if (!fileType || !fileType.mime.startsWith("image/")) return { valid: false, error: "Invalid image file type. Only actual images are allowed" };
+		if (!fileType || !ALLOWED_MIME_TYPES.includes(fileType.mime))
+			return { valid: false, error: "Invalid image file type. Only .jpeg, .png, .gif, and .webp are allowed" };
 
-		const metadata = await sharp(buffer).metadata();
+		let metadata: sharp.Metadata;
+		try {
+			metadata = await sharp(buffer).metadata();
+		} catch {
+			return { valid: false, error: "Invalid or corrupted image file" };
+		}
 
 		// Check image dimensions
 		if (
@@ -49,25 +38,31 @@ export async function validateImage(file: File): Promise<{ valid: boolean; error
 		}
 
 		// Check for inappropriate content
-		// const image = tf.node.decodeImage(buffer, 3) as tf.Tensor3D;
-		// const model = await loadModel();
-		// const predictions = await model.classify(image);
-		// image.dispose();
+		// https://github.com/trafficlunar/api-moderation
+		try {
+			const blob = new Blob([buffer]);
+			const formData = new FormData();
+			formData.append("image", blob);
 
-		// for (const pred of predictions) {
-		// 	if (
-		// 		(pred.className === "Porn" && pred.probability > THRESHOLD) ||
-		// 		(pred.className === "Hentai" && pred.probability > THRESHOLD) ||
-		// 		(pred.className === "Sexy" && pred.probability > THRESHOLD)
-		// 	) {
-		// 		// reject image
-		// 		return { valid: false, error: "Image contains inappropriate content" };
-		// 	}
-		// }
+			const moderationResponse = await fetch("https://api.trafficlunar.net/moderate/image", { method: "POST", body: formData });
+			const result = await moderationResponse.json();
+
+			if (!moderationResponse.ok) {
+				console.error("Moderation API error:", result);
+				return { valid: false, error: "Content moderation check failed", status: 500 };
+			}
+
+			if (result.error) {
+				return { valid: false, error: result.error };
+			}
+		} catch (moderationError) {
+			console.error("Error fetching moderation API:", moderationError);
+			return { valid: false, error: "Moderation API is down", status: 503 };
+		}
+
+		return { valid: true };
 	} catch (error) {
 		console.error("Error validating image:", error);
 		return { valid: false, error: "Failed to process image file.", status: 500 };
 	}
-
-	return { valid: true };
 }