diff --git a/frontend/src/components/discovery/MasonryGrid.vue b/frontend/src/components/discovery/MasonryGrid.vue
index f90a5ce..436b8e9 100644
--- a/frontend/src/components/discovery/MasonryGrid.vue
+++ b/frontend/src/components/discovery/MasonryGrid.vue
@@ -79,9 +79,18 @@ function aspectStyle(item) {
   return { aspectRatio: `${w} / ${h}` }
 }
 
+// Larger rootMargin than the composable default (600px) because the
+// sentinel sits at the BOTTOM of the masonry container, whose height is
+// the MAX of the column heights. A single tall image (long manga page,
+// panorama) in one column pushes the sentinel way past the visible
+// bottom of the SHORTER columns — the user reads the short-column
+// bottoms long before the sentinel comes into view, and load-more
+// fires too late. 2400px ≈ 2-3 screen-heights of pre-emptive trigger,
+// comfortably covering typical tall-image heights. Operator-flagged
+// 2026-05-30.
 useInfiniteScroll(sentinelEl, () => {
   if (props.hasMore && !props.loading) emit('load-more')
-})
+}, { rootMargin: '2400px' })
 </script>
 
 <style scoped>
diff --git a/frontend/src/stores/showcase.js b/frontend/src/stores/showcase.js
index 5e7686e..14055a1 100644
--- a/frontend/src/stores/showcase.js
+++ b/frontend/src/stores/showcase.js
@@ -1,44 +1,96 @@
 import { defineStore } from 'pinia'
 import { ref, computed } from 'vue'
 import { useApi } from '../composables/useApi.js'
-import { useAsyncAction } from '../composables/useAsyncAction.js'
 
-// Operator-confirmed 2026-05-30: instead of one 60-item request, fetch
-// PAGE-sized chunks sequentially so items render as each batch lands
-// rather than blocking on the full 60-item response. Total initial count
-// is unchanged (PAGE * INITIAL_BATCHES = 60). Infinite-scroll also pulls
-// PAGE items per trigger so subsequent appends stay progressive too.
+// Operator-flagged 2026-05-30 (round 2): the per-batch-then-pause cadence
+// from the original sequential-fetch implementation still felt "chunky" —
+// 5 items appear together, then the API round-trip pauses for ~200 ms,
+// then another 5 burst in. To smooth that:
+//   1. Fire all INITIAL_BATCHES fetches IN PARALLEL — collapses the per-
+//      batch round-trip gap (12 concurrent calls return in ~1 RTT instead
+//      of 12 sequential RTTs).
+//   2. Trickle responses into images.value one item at a time with a
+//      small APPEND_DELAY_MS delay between each — the user sees a steady
+//      cadence rather than per-batch bursts.
+// fetchPage (the infinite-scroll path) reuses the same trickle so its
+// 5-item appends also cascade one-by-one instead of popping in together.
 const PAGE = 5
 const INITIAL_BATCHES = 12
+const APPEND_DELAY_MS = 80  // ≈ the MasonryGrid stagger animation (70 ms)
+
+
+function _sleep(ms) { return new Promise(r => setTimeout(r, ms)) }
+
 
 export const useShowcaseStore = defineStore('showcase', () => {
   const api = useApi()
   const images = ref([])
-  const { loading, error, run } = useAsyncAction({ errorAs: 'message' })
+  const loading = ref(false)
+  const error = ref(null)
   const exhausted = ref(false)
   const seen = new Set()
 
-  async function fetchPage() {
-    if (loading.value) return
-    await run(async () => {
-      const body = await api.get('/api/showcase', { params: { limit: PAGE } })
-      const fresh = body.images.filter(i => !seen.has(i.id))
-      if (fresh.length === 0) { exhausted.value = true; return }
-      for (const i of fresh) seen.add(i.id)
-      images.value.push(...fresh)
-    })
+  // Sequence token: every call to loadInitial bumps this. _trickleAppend
+  // bails between items if its captured seq is no longer current — guards
+  // against a fast shuffle / mount-then-shuffle from interleaving two
+  // trickles into the same images.value.
+  let _seq = 0
+
+  async function _trickleAppend(items, mySeq) {
+    for (const item of items) {
+      if (mySeq !== _seq) return
+      if (seen.has(item.id)) continue
+      seen.add(item.id)
+      images.value.push(item)
+      await _sleep(APPEND_DELAY_MS)
+    }
   }
 
-  // Reset state and fetch INITIAL_BATCHES chunks in sequence. Used by
-  // mount, the Shuffle button, and the R-key handler — all want the
-  // same progressive-cascade behavior.
+  // Single batch — used by infinite-scroll appends. Trickles its 5 items
+  // in for the same one-at-a-time cadence as the initial load.
+  async function fetchPage() {
+    if (loading.value) return
+    loading.value = true
+    error.value = null
+    try {
+      const body = await api.get('/api/showcase', { params: { limit: PAGE } })
+      const fresh = (body.images || []).filter(i => !seen.has(i.id))
+      if (fresh.length === 0) { exhausted.value = true; return }
+      await _trickleAppend(fresh, _seq)
+    } catch (e) {
+      error.value = e.message || String(e)
+    } finally {
+      loading.value = false
+    }
+  }
+
+  // Reset state, fire INITIAL_BATCHES fetches in parallel, then trickle
+  // each response's items into images.value at APPEND_DELAY_MS cadence.
+  // Total wall-clock = ~1 RTT (parallel) + N items × APPEND_DELAY_MS;
+  // the user sees a smooth steady stream throughout.
   async function loadInitial() {
+    _seq += 1
+    const mySeq = _seq
     images.value = []
     seen.clear()
     exhausted.value = false
-    for (let i = 0; i < INITIAL_BATCHES; i++) {
-      if (exhausted.value) break
-      await fetchPage()
+    error.value = null
+    loading.value = true
+    try {
+      const fetches = Array.from({ length: INITIAL_BATCHES }, () =>
+        api.get('/api/showcase', { params: { limit: PAGE } }).catch(e => {
+          error.value = error.value || (e.message || String(e))
+          return null
+        })
+      )
+      for (const fetchPromise of fetches) {
+        if (mySeq !== _seq) return
+        const body = await fetchPromise
+        if (body && body.images) await _trickleAppend(body.images, mySeq)
+      }
+      if (mySeq === _seq && images.value.length === 0) exhausted.value = true
+    } finally {
+      if (mySeq === _seq) loading.value = false
     }
   }