diff --git a/.jules/bolt.md b/.jules/bolt.md
index b45f9caa..d62e7fc4 100644
--- a/.jules/bolt.md
+++ b/.jules/bolt.md
@@ -60,3 +60,7 @@ Optimized metric route processing to O(N) by creating a mapping of routes direct
 ## 2026-06-25 - Avoid Map allocations in frontend ERD loops and mutate asyncpg records in-place
 **Learning:** The frontend `snapshotToGraph` iterates over thousands of columns to generate the graph, so repeated lookups and redundant collection assignments increase GC pressure. Backend snapshot column dictionaries are freshly instantiated for the payload, so `add_column_examples` can safely fill missing fields in place.
 **Action:** Reuse existing collections while aggregating relational data, create `Map`/`Set` entries only on first use, and check for missing example fields before calling expensive inference helpers.
+
+## 2024-07-03 - [Frontend Search Optimization]
+**Learning:** React Flow node filtering using array `.flatMap()`, `.join(" ")` and lowercase on the joined string causes significant memory allocations and garbage collection pressure on large graphs.
+**Action:** Avoid mass array creation and string concatenations in hot paths like `useMemo` search filters. Use explicit `for` loops with short-circuiting (`break` and early returns) to minimize unnecessary object allocations and speed up rendering.
diff --git a/frontend/src/App.tsx b/frontend/src/App.tsx
index e99700c5..bd7aaea9 100644
--- a/frontend/src/App.tsx
+++ b/frontend/src/App.tsx
@@ -181,19 +181,29 @@ export default function App() {
   const searchMatchedNodeIds = useMemo(() => {
     if (!normalizedNodeSearch) return new Set<string>();
     const matches = new Set<string>();
+
+    // ⚡ Bolt: Optimized search filtering
+    // Replaced expensive array spreading, flatMap(), and string joining with explicit loops.
+    // Early return (break) on first match avoids processing remaining fields.
+    // This reduces memory allocations and garbage collection overhead during search.
     for (const node of nodes) {
-      const haystack = [
-        node.data.title,
-        node.data.comment ?? "",
-        ...node.data.columns.flatMap((column) => [
-          column.column_name,
-          column.data_type,
-          column.column_comment ?? "",
-        ]),
-      ]
-        .join(" ")
-        .toLocaleLowerCase();
-      if (haystack.includes(normalizedNodeSearch)) {
+      let isMatch = false;
+      const title = node.data.title.toLocaleLowerCase();
+      if (title.includes(normalizedNodeSearch)) {
+        isMatch = true;
+      } else if (node.data.comment && node.data.comment.toLocaleLowerCase().includes(normalizedNodeSearch)) {
+        isMatch = true;
+      } else {
+        for (const column of node.data.columns) {
+          if (column.column_name.toLocaleLowerCase().includes(normalizedNodeSearch) ||
+              column.data_type.toLocaleLowerCase().includes(normalizedNodeSearch) ||
+              (column.column_comment && column.column_comment.toLocaleLowerCase().includes(normalizedNodeSearch))) {
+            isMatch = true;
+            break;
+          }
+        }
+      }
+      if (isMatch) {
         matches.add(node.id);
       }
     }