Add SyNode™ OCR /eyes feature to IntelliNode™ stack

index.html

@@ -686,6 +686,13 @@
             <span>Point Llammy at a YouTube/Instagram clip and it rebuilds the scene in Blender Python</span>
           </div>
         </div>
+        <div class="feature-item">
+          <span class="feature-icon">👁️</span>
+          <div class="feature-text">
+            <strong>SyNode™ OCR — /eyes</strong>
+            <span>Nemotron OCR v2 scans the Blender viewport: extracts text regions, bounding boxes, and confidence scores. Crops feed into LlammyBlenderGGUF (Qwen3-VL) for grounded visual reasoning — Llammy sees what's actually on screen before it responds</span>
+          </div>
+        </div>
       </div>
     </div>
   </section>