diff --git a/lightrag/operate.py b/lightrag/operate.py
index 22e49066..fb46610f 100644
--- a/lightrag/operate.py
+++ b/lightrag/operate.py
@@ -2956,19 +2956,26 @@ async def _apply_token_truncation(
     filtered_entities = []
     filtered_entity_id_to_original = {}
     if entities_context:
-        final_entity_names = {e["entity"] for e in entities_context}
+        entity_name_to_id = {e["entity"]: e["id"] for e in entities_context}
+        final_entity_names = set(entity_name_to_id.keys())
         seen_nodes = set()
         for entity in final_entities:
             name = entity.get("entity_name")
             if name in final_entity_names and name not in seen_nodes:
-                filtered_entities.append(entity)
-                filtered_entity_id_to_original[name] = entity
+                entity_with_id = entity.copy()
+                entity_with_id["id"] = entity_name_to_id.get(name)
+
+                filtered_entities.append(entity_with_id)
+                filtered_entity_id_to_original[name] = entity_with_id
                 seen_nodes.add(name)
 
     filtered_relations = []
     filtered_relation_id_to_original = {}
     if relations_context:
-        final_relation_pairs = {(r["entity1"], r["entity2"]) for r in relations_context}
+        relation_pair_to_id = {
+            (r["entity1"], r["entity2"]): r["id"] for r in relations_context
+        }
+        final_relation_pairs = set(relation_pair_to_id.keys())
         seen_edges = set()
         for relation in final_relations:
             src, tgt = relation.get("src_id"), relation.get("tgt_id")
@@ -2977,8 +2984,11 @@ async def _apply_token_truncation(
 
             pair = (src, tgt)
             if pair in final_relation_pairs and pair not in seen_edges:
-                filtered_relations.append(relation)
-                filtered_relation_id_to_original[pair] = relation
+                relation_with_id = relation.copy()
+                relation_with_id["id"] = relation_pair_to_id.get(pair)
+
+                filtered_relations.append(relation_with_id)
+                filtered_relation_id_to_original[pair] = relation_with_id
                 seen_edges.add(pair)
 
     return {
diff --git a/lightrag/utils.py b/lightrag/utils.py
index 7692cb2a..43b8c0a3 100644
--- a/lightrag/utils.py
+++ b/lightrag/utils.py
@@ -2744,6 +2744,7 @@ def _convert_to_user_format(
             # Use original database data
             formatted_entities.append(
                 {
+                    "id": original_entity.get("id", "unknown"),
                     "entity_name": original_entity.get("entity_name", entity_name),
                     "entity_type": original_entity.get("entity_type", "UNKNOWN"),
                     "description": original_entity.get("description", ""),
@@ -2756,6 +2757,7 @@ def _convert_to_user_format(
             # Fallback to LLM context data (for backward compatibility)
             formatted_entities.append(
                 {
+                    "id": entity.get("id", "unknown"),
                     "entity_name": entity_name,
                     "entity_type": entity.get("type", "UNKNOWN"),
                     "description": entity.get("description", ""),
@@ -2781,6 +2783,7 @@ def _convert_to_user_format(
             # Use original database data
             formatted_relationships.append(
                 {
+                    "id": original_relation.get("id", "unknown"),
                     "src_id": original_relation.get("src_id", entity1),
                     "tgt_id": original_relation.get("tgt_id", entity2),
                     "description": original_relation.get("description", ""),
@@ -2795,6 +2798,7 @@ def _convert_to_user_format(
             # Fallback to LLM context data (for backward compatibility)
             formatted_relationships.append(
                 {
+                    "id": relation.get("id", "unknown"),
                     "src_id": entity1,
                     "tgt_id": entity2,
                     "description": relation.get("description", ""),
@@ -2810,6 +2814,7 @@ def _convert_to_user_format(
     formatted_chunks = []
     for i, chunk in enumerate(final_chunks):
         chunk_data = {
+            "id": chunk.get("id", "unknown"),
             "content": chunk.get("content", ""),
             "file_path": chunk.get("file_path", "unknown_source"),
             "chunk_id": chunk.get("chunk_id", ""),