diff --git a/.jules/bolt.md b/.jules/bolt.md
new file mode 100644
index 00000000..0205bf85
--- /dev/null
+++ b/.jules/bolt.md
@@ -0,0 +1,3 @@
+## 2024-06-25 - Parallelizing State-Dependent Array Fetches
+**Learning:** Sequential network calls inside a loop (`for...of`) create cumulative latency, especially when collecting configuration items across multiple potential reverse proxy paths. It's safe to parallelize these requests if the state side-effects (e.g., updating `lastStatus`, `lastError`, `lastRaw`) depend strictly on the sequence order.
+**Action:** Replace sequential loops with `await Promise.all(paths.map(fetchFn))`. Then, iterate through the resolved results array (`for (const r of results)`) to process side effects. This guarantees order-dependent logic remains identical while shifting the I/O bottleneck from serial to parallel execution.
diff --git a/src/lib/forge-openai-surface.ts b/src/lib/forge-openai-surface.ts
index 93415262..9c58541e 100644
--- a/src/lib/forge-openai-surface.ts
+++ b/src/lib/forge-openai-surface.ts
@@ -118,8 +118,11 @@ export async function collectZimaOSV1ModelEntries(
   let lastError: string | undefined;
   let lastRaw: unknown = null;
 
-  for (const path of V1_MODELS_PATHS) {
-    const r = await fetchZimaOSJson(email, path);
+  const results = await Promise.all(
+    V1_MODELS_PATHS.map(path => fetchZimaOSJson(email, path))
+  );
+
+  for (const r of results) {
     lastStatus = r.status;
     lastError = r.error;
     lastRaw = r.data;