diff --git a/libs/langchain/langchain/vectorstores/deeplake.py b/libs/langchain/langchain/vectorstores/deeplake.py
index b6223772ab0..f8fa6a8fc89 100644
--- a/libs/langchain/langchain/vectorstores/deeplake.py
+++ b/libs/langchain/langchain/vectorstores/deeplake.py
@@ -413,7 +413,7 @@ class DeepLake(VectorStore):
             distance_metric=distance_metric,
             filter=filter,
             exec_option=exec_option,
-            return_tensors=["embedding", "metadata", "text", "id"],
+            return_tensors=["embedding", "metadata", "text", self._id_tensor_name],
             deep_memory=deep_memory,
         )
 
diff --git a/libs/langchain/tests/integration_tests/vectorstores/test_deeplake.py b/libs/langchain/tests/integration_tests/vectorstores/test_deeplake.py
index a666329cce8..5e943d5cf4e 100644
--- a/libs/langchain/tests/integration_tests/vectorstores/test_deeplake.py
+++ b/libs/langchain/tests/integration_tests/vectorstores/test_deeplake.py
@@ -259,3 +259,25 @@ def test_add_texts(deeplake_datastore: DeepLake) -> None:
             texts=texts,
             metada=metadatas,
         )
+
+
+def test_ids_backwards_compatibility() -> None:
+    """Test that ids are backwards compatible."""
+    db = DeepLake(
+        dataset_path="mem://test_path",
+        embedding_function=FakeEmbeddings(),
+        tensor_params=[
+            {"name": "ids", "htype": "text"},
+            {"name": "text", "htype": "text"},
+            {"name": "embedding", "htype": "embedding"},
+            {"name": "metadata", "htype": "json"},
+        ],
+    )
+    db.vectorstore.add(
+        ids=["1", "2", "3"],
+        text=["foo", "bar", "baz"],
+        embedding=FakeEmbeddings().embed_documents(["foo", "bar", "baz"]),
+        metadata=[{"page": str(i)} for i in range(3)],
+    )
+    output = db.similarity_search("foo", k=1)
+    assert len(output) == 1