diff --git a/backend/packages/harness/deerflow/runtime/events/store/memory.py b/backend/packages/harness/deerflow/runtime/events/store/memory.py
index cf70e1cdf..a2bb54819 100644
--- a/backend/packages/harness/deerflow/runtime/events/store/memory.py
+++ b/backend/packages/harness/deerflow/runtime/events/store/memory.py
@@ -6,6 +6,7 @@ since all mutations happen within the same event loop).
 
 from __future__ import annotations
 
+import bisect
 from datetime import UTC, datetime
 
 from deerflow.runtime.events.store.base import RunEventStore
@@ -13,7 +14,11 @@ from deerflow.runtime.events.store.base import RunEventStore
 
 class MemoryRunEventStore(RunEventStore):
     def __init__(self) -> None:
-        self._events: dict[str, list[dict]] = {}  # thread_id -> sorted event list
+        self._events: dict[str, list[dict]] = {}  # thread_id -> seq-sorted event list
+        # Messages-only projection of ``_events`` (same dict objects, no copies),
+        # kept in seq order so message pagination is O(log m + page) via bisect
+        # instead of re-scanning every event on each request.
+        self._messages: dict[str, list[dict]] = {}  # thread_id -> seq-sorted message list
         self._seq_counters: dict[str, int] = {}  # thread_id -> last assigned seq
 
     def _next_seq(self, thread_id: str) -> int:
@@ -45,6 +50,8 @@ class MemoryRunEventStore(RunEventStore):
             "created_at": created_at or datetime.now(UTC).isoformat(),
         }
         self._events.setdefault(thread_id, []).append(record)
+        if category == "message":
+            self._messages.setdefault(thread_id, []).append(record)
         return record
 
     async def put(
@@ -76,18 +83,20 @@ class MemoryRunEventStore(RunEventStore):
         return results
 
     async def list_messages(self, thread_id, *, limit=50, before_seq=None, after_seq=None):
-        all_events = self._events.get(thread_id, [])
-        messages = [e for e in all_events if e["category"] == "message"]
+        # ``messages`` is messages-only and seq-sorted, so the seq window is a
+        # contiguous slice located with bisect (O(log m)) rather than a full scan.
+        messages = self._messages.get(thread_id, [])
 
         if before_seq is not None:
-            messages = [e for e in messages if e["seq"] < before_seq]
-            # Take the last `limit` records
-            return messages[-limit:]
+            # Records with seq < before_seq, then the last `limit` of them.
+            hi = bisect.bisect_left(messages, before_seq, key=lambda e: e["seq"])
+            return messages[max(0, hi - limit) : hi]
         elif after_seq is not None:
-            messages = [e for e in messages if e["seq"] > after_seq]
-            return messages[:limit]
+            # Records with seq > after_seq, then the first `limit` of them.
+            lo = bisect.bisect_right(messages, after_seq, key=lambda e: e["seq"])
+            return messages[lo : lo + limit]
         else:
-            # Return the latest `limit` records, ascending
+            # Return the latest `limit` records, ascending.
             return messages[-limit:]
 
     async def list_events(self, thread_id, run_id, *, event_types=None, limit=500):
@@ -110,11 +119,11 @@ class MemoryRunEventStore(RunEventStore):
             return filtered[-limit:] if len(filtered) > limit else filtered
 
     async def count_messages(self, thread_id):
-        all_events = self._events.get(thread_id, [])
-        return sum(1 for e in all_events if e["category"] == "message")
+        return len(self._messages.get(thread_id, []))
 
     async def delete_by_thread(self, thread_id):
         events = self._events.pop(thread_id, [])
+        self._messages.pop(thread_id, None)
         self._seq_counters.pop(thread_id, None)
         return len(events)
 
@@ -125,4 +134,6 @@ class MemoryRunEventStore(RunEventStore):
         remaining = [e for e in all_events if e["run_id"] != run_id]
         removed = len(all_events) - len(remaining)
         self._events[thread_id] = remaining
+        # Keep the message projection in lockstep (same surviving dict objects).
+        self._messages[thread_id] = [e for e in remaining if e["category"] == "message"]
         return removed
diff --git a/backend/tests/test_run_event_store.py b/backend/tests/test_run_event_store.py
index 17b796af7..7c409c711 100644
--- a/backend/tests/test_run_event_store.py
+++ b/backend/tests/test_run_event_store.py
@@ -122,6 +122,26 @@ class TestListMessages:
         messages = await store.list_messages("t1", limit=3)
         assert [m["seq"] for m in messages] == [8, 9, 10]
 
+    @pytest.mark.anyio
+    async def test_pagination_with_interleaved_trace_events(self, store):
+        # Messages and non-message events interleave, so message seqs are
+        # non-contiguous (1, 3, 5, 7, 9). Seq-window pagination must still be
+        # correct over the messages-only projection, including when the cursor
+        # lands in a gap or exactly on a message seq (exclusive bound).
+        for i in range(10):
+            category = "message" if i % 2 == 0 else "trace"
+            await store.put(thread_id="t1", run_id="r1", event_type="e", category=category, content=str(i))
+
+        assert [m["seq"] for m in await store.list_messages("t1")] == [1, 3, 5, 7, 9]
+        # before_seq in a gap: seq < 6 -> [1, 3, 5], last 2
+        assert [m["seq"] for m in await store.list_messages("t1", before_seq=6, limit=2)] == [3, 5]
+        # before_seq on a message seq is exclusive: seq < 5 -> [1, 3]
+        assert [m["seq"] for m in await store.list_messages("t1", before_seq=5, limit=5)] == [1, 3]
+        # after_seq in a gap: seq > 4 -> [5, 7, 9], first 2
+        assert [m["seq"] for m in await store.list_messages("t1", after_seq=4, limit=2)] == [5, 7]
+        # after_seq on a message seq is exclusive: seq > 5 -> [7, 9]
+        assert [m["seq"] for m in await store.list_messages("t1", after_seq=5, limit=5)] == [7, 9]
+
 
 # -- list_events --