diff --git a/package.json b/package.json
index 0939c02b..b2934fdc 100644
--- a/package.json
+++ b/package.json
@@ -113,7 +113,8 @@
     "undici": "^5.28.5",
     "vaul": "^1.1.2",
     "vercel": "^44.7.3",
-    "zod": "^4.0.17"
+    "zod": "^4.0.17",
+    "cheerio": "^1.1.2"
   },
   "devDependencies": {
     "@eslint/js": "^9.33.0",
diff --git a/src/lib/search-service.ts b/src/lib/search-service.ts
index 8b5d9157..49f6c9e9 100644
--- a/src/lib/search-service.ts
+++ b/src/lib/search-service.ts
@@ -1,7 +1,7 @@
 import * as Sentry from '@sentry/react';
 import { withTimeout } from './ai-utils';
 import { createTokenBucketRateLimiter } from './rate-limiter';
-
+import * as cheerio from 'cheerio';
 const { logger } = Sentry;
 
 // Simple in-memory rate limiter per session (client-side)
@@ -316,12 +316,14 @@ export class BraveSearchService {
   }
 
   private extractTextContent(html: string): string {
-    return html
-      .replace(/<script[^>]*>.*?<\/script>/gi, '')
-      .replace(/<style[^>]*>.*?<\/style>/gi, '')
-      .replace(/<[^>]*>/g, ' ')
-      .replace(/\s+/g, ' ')
-      .trim();
+    // Use Cheerio to parse HTML and remove <script> and <style> tags
+    const $ = cheerio.load(html);
+    $('script').remove();
+    $('style').remove();
+    // Get the text content
+    const text = $.root().text();
+    // Collapse whitespace and trim
+    return text.replace(/\s+/g, ' ').trim();
   }
 }