diff --git a/__init__.py b/__init__.py
index be00731..f024196 100644
--- a/__init__.py
+++ b/__init__.py
@@ -162,6 +162,7 @@ NODE_CONFIG = {
     "TorchCompileControlNet": {"class": TorchCompileControlNet, "name": "TorchCompileControlNet"},
     "PatchModelPatcherOrder": {"class": PatchModelPatcherOrder, "name": "Patch Model Patcher Order"},
     "TorchCompileLTXModel": {"class": TorchCompileLTXModel, "name": "TorchCompileLTXModel"},
+    "PathchSageAttentionKJ": {"class": PathchSageAttentionKJ, "name": "Pathch Sage Attention KJ"},
 
     #instance diffusion
     "CreateInstanceDiffusionTracking": {"class": CreateInstanceDiffusionTracking},
diff --git a/nodes/model_optimization_nodes.py b/nodes/model_optimization_nodes.py
index 33b5d52..8593bf6 100644
--- a/nodes/model_optimization_nodes.py
+++ b/nodes/model_optimization_nodes.py
@@ -114,6 +114,24 @@ class BaseLoaderKJ:
                 disable_weight_init.Linear = BaseLoaderKJ.original_linear
                 BaseLoaderKJ.cublas_patched = False
 
+class PathchSageAttentionKJ(BaseLoaderKJ):
+    @classmethod
+    def INPUT_TYPES(s):
+        return {"required": {
+            "model": ("MODEL",),
+            "sage_attention": (["disabled", "auto", "sageattn_qk_int8_pv_fp16_cuda", "sageattn_qk_int8_pv_fp16_triton", "sageattn_qk_int8_pv_fp8_cuda"], {"default": False, "tooltip": "Patch comfy attention to use sageattn."}),
+        }}
+
+    RETURN_TYPES = ("MODEL", )
+    FUNCTION = "patch"
+    DESCRIPTION = "Experimental node for patching attention mode."
+    EXPERIMENTAL = True
+    CATEGORY = "KJNodes/experimental"
+
+    def patch(self, model, sage_attention):
+        self._patch_modules(False, sage_attention)
+        return model,
+ 
 class CheckpointLoaderKJ(BaseLoaderKJ):
     @classmethod
     def INPUT_TYPES(s):