diff --git a/config.json b/config.json
index a16be7b2..85646de3 100644
--- a/config.json
+++ b/config.json
@@ -23,5 +23,5 @@
 			"files_file"
 		]
 	},
-	"instance_version": "6.11.22"
+	"instance_version": "6.12.8"
 }
diff --git a/src/main.py b/src/main.py
index e9e99d1a..baaa52c6 100644
--- a/src/main.py
+++ b/src/main.py
@@ -200,6 +200,8 @@ def run_pipeline_from_api(response: Response, request: Request):
         # pipeline_preset = state["pipeline_preset"]
         if g.pipeline_running:
             return {"result": "pipeline is already running. Please wait until it's finished."}
+
+        time.sleep(5)  # delay to init layers
         run_btn_clicked()
         return {"result": "pipeline was successfully processed"}
     except Exception as e:
diff --git a/src/scripts/api_request.py b/src/scripts/api_request.py
new file mode 100644
index 00000000..26de1a1c
--- /dev/null
+++ b/src/scripts/api_request.py
@@ -0,0 +1,28 @@
+# Example on how to send API requests to ML Pipelines app from a script
+# Available endpoints:
+# 1. run_pipeline - runs a pipeline with currently selected layers on the scene
+# 2. get_pipeline_status - returns the status of the pipeline (e.g "Pipeline status is 21/50")
+
+import os
+
+import supervisely as sly
+from dotenv import load_dotenv
+
+load_dotenv(os.path.expanduser("~/supervisely.env"))
+load_dotenv("local.env")
+
+team_id = sly.env.team_id()
+workspace_id = sly.env.workspace_id()
+
+api: sly.Api = sly.Api.from_env()
+
+task_id = 68139
+
+is_ready = api.app.is_ready_for_api_calls(task_id)
+
+# is_ready = api.app.wait_until_ready_for_api_calls(task_id)
+# if is_ready:
+api.task.send_request(task_id=task_id, method="run_pipeline", data={})
+
+# status = api.task.send_request(task_id=task_id, method="get_pipeline_status", data={})
+# print(status)
diff --git a/src/scripts/auto_start.py b/src/scripts/auto_start.py
new file mode 100644
index 00000000..015db18e
--- /dev/null
+++ b/src/scripts/auto_start.py
@@ -0,0 +1,57 @@
+# AutoStart ML Pipelines from preset file
+
+import os
+
+import supervisely as sly
+from dotenv import load_dotenv
+
+load_dotenv(os.path.expanduser("~/supervisely.env"))
+load_dotenv("local.env")
+
+team_id = sly.env.team_id()
+workspace_id = sly.env.workspace_id()
+
+api: sly.Api = sly.Api.from_env()
+
+agent_id = 452  # 359
+app_slug = "supervisely-ecosystem/data-nodes"
+
+# preset_path = "/data-nodes/presets/images/preset.json"
+preset_path = "/data-nodes/presets/images/api_pipeline.json"
+
+module_id = api.app.get_ecosystem_module_id(app_slug)
+module_info = api.app.get_ecosystem_module_info(module_id)
+
+
+params = {"modalityType": "images", "slyFile": preset_path}
+
+# app_params = {
+#     "agent_id": agent_id,
+#     # "app_id": 0,
+#     "module_id": module_id,
+#     "workspace_id": workspace_id,
+#     "description": "Start ML Pipelines from py",
+#     "task_name": "ML Pipelines",
+#     "params": {"autostart": False, **params},
+#     "app_version": None,
+#     "is_branch": False,
+# }
+
+app_params = {
+    "agent_id": agent_id,
+    # "app_id": 0,
+    "module_id": module_id,
+    "workspace_id": workspace_id,
+    "description": "Start ML Pipelines from API",
+    "task_name": "ML Pipelines",
+    "params": {"autostart": False, **params},
+    "app_version": "run-from-api",
+    "is_branch": True,
+}
+
+session_info = api.app.start(**app_params)
+
+# Run pipeline
+is_ready = api.app.wait_until_ready_for_api_calls(session_info.task_id)
+if is_ready:
+    api.task.send_request(task_id=session_info.task_id, method="run_pipeline", data={})