From d362fa75371d3b8d5b358e3cd2767f4b4258d560 Mon Sep 17 00:00:00 2001
From: "gpt-engineer-app[bot]"
 <159125892+gpt-engineer-app[bot]@users.noreply.github.com>
Date: Wed, 29 Oct 2025 20:45:21 +0000
Subject: [PATCH] feat: Improve bot detection for Vercel OG preview

---
 api/ssrOG.ts | 43 ++++++++++++++++++++++++++++++++++++-------
 1 file changed, 36 insertions(+), 7 deletions(-)

diff --git a/api/ssrOG.ts b/api/ssrOG.ts
index d225fb4f..40e85b01 100644
--- a/api/ssrOG.ts
+++ b/api/ssrOG.ts
@@ -37,7 +37,23 @@ const SOCIAL_BOTS = {
   'slurp': 'yahoo',
   'duckduckbot': 'duckduckgo',
   'baiduspider': 'baidu',
-  'yandexbot': 'yandex'
+  'yandexbot': 'yandex',
+  // Headless browsers & crawlers
+  'headless': 'headless-browser',
+  'chrome-lighthouse': 'lighthouse',
+  'puppeteer': 'puppeteer',
+  'playwright': 'playwright',
+  'selenium': 'selenium',
+  'phantomjs': 'phantomjs',
+  // Vercel & deployment platforms
+  'vercel': 'vercel',
+  'vercel-screenshot': 'vercel',
+  'prerender': 'prerender',
+  // Generic crawler patterns
+  'bot': 'generic-bot',
+  'crawler': 'generic-crawler',
+  'spider': 'generic-spider',
+  'scraper': 'generic-scraper'
 };
 
 interface BotDetection {
@@ -229,20 +245,35 @@ export default async function handler(req: VercelRequest, res: VercelResponse) {
     const fullUrl = `https://${req.headers.host}${req.url}`;
     const pathname = new URL(fullUrl).pathname;
     
-    console.log(`[SSR-OG] ${req.method} ${pathname} | UA: ${userAgent.substring(0, 60)}`);
-    
     // Bot detection
     const botDetection = detectBot(userAgent);
     
+    // Enhanced logging
+    if (botDetection.isBot) {
+      console.log(`[SSR-OG] ✅ Bot detected: ${botDetection.platform} | ${req.method} ${pathname}`);
+      console.log(`[SSR-OG] Full UA: ${userAgent}`);
+    } else {
+      // Log undetected potential bots for debugging
+      const looksLikeBot = !userAgent.includes('Mozilla') || 
+                           userAgent.includes('http') || 
+                           userAgent.length < 50;
+      
+      if (looksLikeBot) {
+        console.warn(`[SSR-OG] ⚠️ Possible undetected bot | ${req.method} ${pathname}`);
+        console.warn(`[SSR-OG] Full UA: ${userAgent}`);
+      } else {
+        console.log(`[SSR-OG] Regular user | ${req.method} ${pathname} | UA: ${userAgent.substring(0, 60)}...`);
+      }
+    }
+    
     // Read the built index.html
     const htmlPath = join(process.cwd(), 'dist', 'index.html');
     let html = readFileSync(htmlPath, 'utf-8');
     
     if (botDetection.isBot) {
-      console.log(`[SSR-OG] Bot detected: ${botDetection.platform}`);
-      
       // Fetch page-specific data
       const pageData = await getPageData(pathname, fullUrl);
+      console.log(`[SSR-OG] Generated OG tags: ${pageData.title}`);
       
       // Generate and inject OG tags
       const ogTags = generateOGTags(pageData);
@@ -250,8 +281,6 @@ export default async function handler(req: VercelRequest, res: VercelResponse) {
       
       res.setHeader('X-Bot-Platform', botDetection.platform || 'unknown');
       res.setHeader('X-SSR-Modified', 'true');
-    } else {
-      console.log('[SSR-OG] Regular user - serving original HTML');
     }
     
     res.setHeader('Content-Type', 'text/html; charset=utf-8');