diff --git a/docs/advanced_features/vlm_query.ipynb b/docs/advanced_features/vlm_query.ipynb
index f5497f5b7..dd35f5eae 100644
--- a/docs/advanced_features/vlm_query.ipynb
+++ b/docs/advanced_features/vlm_query.ipynb
@@ -64,6 +64,8 @@
     "\n",
     "nest_asyncio.apply()\n",
     "\n",
+    "import sglang.test.doc_patch  # noqa: F401\n",
+    "\n",
     "model_path = \"Qwen/Qwen2.5-VL-3B-Instruct\"\n",
     "chat_template = \"qwen2-vl\"\n",
     "example_image_url = \"https://raw.githubusercontent.com/sgl-project/sglang/main/examples/assets/example_image.png\""
diff --git a/docs/basic_usage/offline_engine_api.ipynb b/docs/basic_usage/offline_engine_api.ipynb
index 9c03e90a7..fe8a9e304 100644
--- a/docs/basic_usage/offline_engine_api.ipynb
+++ b/docs/basic_usage/offline_engine_api.ipynb
@@ -66,7 +66,7 @@
     "import asyncio\n",
     "\n",
     "import sglang as sgl\n",
-    "import sglang.test.doc_patch\n",
+    "import sglang.test.doc_patch  # noqa: F401\n",
     "from sglang.utils import async_stream_and_merge, stream_and_merge\n",
     "\n",
     "llm = sgl.Engine(model_path=\"qwen/qwen2.5-0.5b-instruct\")"
diff --git a/python/sglang/test/doc_patch.py b/python/sglang/test/doc_patch.py
index 0febb295c..503ce86e7 100644
--- a/python/sglang/test/doc_patch.py
+++ b/python/sglang/test/doc_patch.py
@@ -3,6 +3,7 @@ Do some monkey patch to make the documentation compilation faster and more relia
 
 - Avoid port conflicts
 - Reduce the server launch time
+- Limit GPU memory usage to allow multiple servers on the same machine
 """
 
 import weakref
@@ -26,6 +27,9 @@ def patched_post_init(self):
         self.max_running_requests = DEFAULT_MAX_RUNNING_REQUESTS
     if self.max_total_tokens is None:
         self.max_total_tokens = DEFAULT_MAX_TOTAL_TOKENS
+    # Disable CUDA graphs to avoid memory spikes during capture.
+    # Notebooks only run a few sample requests, so perf is not critical.
+    self.disable_cuda_graph = True
     self.cuda_graph_max_bs = 4
 
 
@@ -47,6 +51,7 @@ def launch_server_cmd(command: str, host: str = "0.0.0.0", port: int = None):
     extra_flags = (
         f"--max-running-requests {DEFAULT_MAX_RUNNING_REQUESTS} "
         f"--max-total-tokens {DEFAULT_MAX_TOTAL_TOKENS} "
+        f"--disable-cuda-graph "
         f"--cuda-graph-max-bs 4"
     )