diff --git a/src/memos/graph_dbs/polardb.py b/src/memos/graph_dbs/polardb.py
index d2d69c768..74dd38fc1 100644
--- a/src/memos/graph_dbs/polardb.py
+++ b/src/memos/graph_dbs/polardb.py
@@ -1553,7 +1553,7 @@ def search_by_fulltext(
         """
 
         params = [tsquery_string, tsquery_string]
-
+        logger.info(f"[search_by_fulltext] query: {query}, params: {params}")
         conn = self._get_connection()
         try:
             with conn.cursor() as cursor:
diff --git a/src/memos/memories/textual/tree_text_memory/retrieve/recall.py b/src/memos/memories/textual/tree_text_memory/retrieve/recall.py
index 7ac274a62..5dfbde704 100644
--- a/src/memos/memories/textual/tree_text_memory/retrieve/recall.py
+++ b/src/memos/memories/textual/tree_text_memory/retrieve/recall.py
@@ -101,13 +101,27 @@ def retrieve(
                     user_name=user_name,
                     search_filter=id_filter,
                 )
+            if use_fast_graph:
+                future_fulltext = executor.submit(
+                    self._fulltext_recall,
+                    query_words=parsed_goal.keys or [],
+                    memory_scope=memory_scope,
+                    top_k=top_k,
+                    search_filter=search_filter,
+                    search_priority=search_priority,
+                    user_name=user_name,
+                )
 
             graph_results = future_graph.result()
             vector_results = future_vector.result()
             bm25_results = future_bm25.result() if self.use_bm25 else []
+            fulltext_results = future_fulltext.result() if use_fast_graph else []
 
         # Merge and deduplicate by ID
-        combined = {item.id: item for item in graph_results + vector_results + bm25_results}
+        combined = {
+            item.id: item
+            for item in graph_results + vector_results + bm25_results + fulltext_results
+        }
 
         return list(combined.values())
 
@@ -404,3 +418,55 @@ def _bm25_recall(
         )
 
         return [TextualMemoryItem.from_dict(n) for n in bm25_results]
+
+    def _fulltext_recall(
+        self,
+        query_words: list[str],
+        memory_scope: str,
+        top_k: int = 20,
+        max_num: int = 5,
+        status: str = "activated",
+        cube_name: str | None = None,
+        search_filter: dict | None = None,
+        search_priority: dict | None = None,
+        user_name: str | None = None,
+    ):
+        """Perform fulltext-based retrieval.
+        Args:
+            query_words: list of query words
+            memory_scope: memory scope
+            top_k: top k results
+            max_num: max number of query words
+            status: status
+            cube_name: cube name
+            search_filter: search filter
+            search_priority: search priority
+            user_name: user name
+        Returns:
+            list of TextualMemoryItem
+        """
+        if not query_words:
+            return []
+        logger.info(f"[FULLTEXT] query_words: {query_words}")
+        all_hits = self.graph_store.search_by_fulltext(
+            query_words=query_words,
+            top_k=top_k,
+            status=status,
+            scope=memory_scope,
+            cube_name=cube_name,
+            search_filter=search_priority,
+            filter=search_filter,
+            user_name=user_name,
+        )
+        if not all_hits:
+            return []
+
+        # merge and deduplicate
+        unique_ids = {r["id"] for r in all_hits if r.get("id")}
+        node_dicts = (
+            self.graph_store.get_nodes(
+                list(unique_ids), include_embedding=False, cube_name=cube_name, user_name=user_name
+            )
+            or []
+        )
+        return [TextualMemoryItem.from_dict(n) for n in node_dicts]