diff --git a/.vscode/settings.json b/.vscode/settings.json
index 7b281e81b..fc09d6a12 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -11,5 +11,6 @@
   "python.analysis.extraPaths": [
     "packages/server/engine-lib/rocketlib-python/lib",
     "packages/ai/src"
-  ]
+  ],
+  "DockerRun.DisableDockerrc": true
 }
diff --git a/nodes/src/nodes/ml_sklearn/IGlobal.py b/nodes/src/nodes/ml_sklearn/IGlobal.py
new file mode 100644
index 000000000..78541ef5e
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/IGlobal.py
@@ -0,0 +1,42 @@
+# =============================================================================
+# MIT License
+# Copyright (c) 2026 RocketRide Contributors
+# =============================================================================
+import os
+from rocketlib import IGlobalBase, OPEN_MODE, warning
+from ai.common.config import Config
+
+
+class IGlobal(IGlobalBase):
+    """Global state for the ml_sklearn node — holds the loaded sklearn model."""
+
+    preprocessor: object = None
+
+    def validateConfig(self):
+        """Validate that scikit-learn and joblib are available."""
+        try:
+            from depends import depends
+
+            requirements = os.path.dirname(os.path.realpath(__file__)) + '/requirements.txt'
+            depends(requirements)
+        except Exception as e:
+            warning(str(e))
+
+    def beginGlobal(self):
+        """Load the sklearn model at runtime startup."""
+        if self.IEndpoint.endpoint.openMode == OPEN_MODE.CONFIG:
+            pass
+        else:
+            from depends import depends
+
+            requirements = os.path.dirname(os.path.realpath(__file__)) + '/requirements.txt'
+            depends(requirements)
+
+            from .code import PreProcessor
+
+            config = Config.getNodeConfig(self.glb.logicalType, self.glb.connConfig)
+            self.preprocessor = PreProcessor(config)
+
+    def endGlobal(self):
+        """Release the sklearn model."""
+        self.preprocessor = None
diff --git a/nodes/src/nodes/ml_sklearn/IInstance.py b/nodes/src/nodes/ml_sklearn/IInstance.py
new file mode 100644
index 000000000..b3ef8bbe3
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/IInstance.py
@@ -0,0 +1,28 @@
+# =============================================================================
+# MIT License
+# Copyright (c) 2026 RocketRide Contributors
+# =============================================================================
+import copy
+from rocketlib import IInstanceBase, Entry
+from ai.common.schema import Answer
+from .IGlobal import IGlobal
+
+
+class IInstance(IInstanceBase):
+    IGlobal: IGlobal
+
+    def open(self, obj: Entry):
+        pass
+
+    def _process(self, text: str) -> str:
+        if self.IGlobal.preprocessor is not None:
+            return self.IGlobal.preprocessor.process(text)
+        return text
+
+    def writeAnswers(self, answer: Answer):
+        """Process answer text through sklearn model and forward downstream."""
+        answer = copy.deepcopy(answer)
+        text = answer.getText() if answer else ''
+        result = self._process(text)
+        answer.setText(result)
+        self.instance.writeAnswers(answer)
diff --git a/nodes/src/nodes/ml_sklearn/README.md b/nodes/src/nodes/ml_sklearn/README.md
new file mode 100644
index 000000000..aa9d356e9
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/README.md
@@ -0,0 +1,19 @@
+# ML Sklearn Prediction Node
+
+This node performs predictions using a trained scikit-learn model.
+
+## Input
+
+- text (number as string)
+
+## Output
+
+- text (predicted value as string)
+
+## Example
+
+Input:
+250
+
+Output:
+3.5
diff --git a/nodes/src/nodes/ml_sklearn/__init__.py b/nodes/src/nodes/ml_sklearn/__init__.py
new file mode 100644
index 000000000..70eda8d6b
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/__init__.py
@@ -0,0 +1,4 @@
+from .IGlobal import IGlobal
+from .IInstance import IInstance
+
+__all__ = ['IGlobal', 'IInstance']
diff --git a/nodes/src/nodes/ml_sklearn/code.py b/nodes/src/nodes/ml_sklearn/code.py
new file mode 100644
index 000000000..e194def99
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/code.py
@@ -0,0 +1,35 @@
+# =============================================================================
+# MIT License
+# Copyright (c) 2026 RocketRide Contributors
+# =============================================================================
+import os
+import joblib
+
+
+class PreProcessor:
+    """Wraps a scikit-learn model/pipeline for text inference."""
+
+    def __init__(self, config: dict):
+        model_path = config.get('model_path', '')
+
+        self._model = None
+        if model_path and os.path.exists(model_path):
+            self._model = joblib.load(model_path)
+
+    def process(self, text: str) -> str:
+        """
+        Run sklearn inference on input text.
+        Returns prediction as string.
+        Falls back to original text if no model loaded or inference fails.
+        """
+        if self._model is None:
+            return text
+
+        try:
+            value = float(text)
+            prediction = self._model.predict([[value]])
+            return str(prediction[0])
+        except (ValueError, TypeError):
+            return text
+        except Exception:
+            return text
diff --git a/nodes/src/nodes/ml_sklearn/create_sample_model.py b/nodes/src/nodes/ml_sklearn/create_sample_model.py
new file mode 100644
index 000000000..d42a268d5
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/create_sample_model.py
@@ -0,0 +1,23 @@
+# =============================================================================
+# Run this once to generate a sample model.pkl for testing:
+#   python nodes/src/nodes/ml_sklearn/create_sample_model.py
+# =============================================================================
+import joblib
+import numpy as np
+from sklearn.linear_model import LinearRegression
+
+# Simple model: y = 2x + 1
+X = np.array([[1.0], [2.0], [3.0], [4.0], [5.0]])
+y = np.array([3.0, 5.0, 7.0, 9.0, 11.0])
+
+model = LinearRegression()
+model.fit(X, y)
+
+import os
+
+save_path = os.path.join(os.path.dirname(__file__), 'model.pkl')
+joblib.dump(model, save_path)
+
+print(f'✅ Model saved: {save_path}')
+print(f'✅ predict(4.0) = {model.predict([[4.0]])[0]:.1f}  (expected 9.0)')
+print(f'✅ predict(6.0) = {model.predict([[6.0]])[0]:.1f}  (expected 13.0)')
diff --git a/nodes/src/nodes/ml_sklearn/model.pkl b/nodes/src/nodes/ml_sklearn/model.pkl
new file mode 100644
index 000000000..b02586283
Binary files /dev/null and b/nodes/src/nodes/ml_sklearn/model.pkl differ
diff --git a/nodes/src/nodes/ml_sklearn/requirements.txt b/nodes/src/nodes/ml_sklearn/requirements.txt
new file mode 100644
index 000000000..2fb0f3743
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/requirements.txt
@@ -0,0 +1,3 @@
+scikit-learn>=1.3.0
+joblib>=1.3.0
+numpy>=1.24.0
diff --git a/nodes/src/nodes/ml_sklearn/services.json b/nodes/src/nodes/ml_sklearn/services.json
new file mode 100644
index 000000000..1f04676f3
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/services.json
@@ -0,0 +1,72 @@
+{
+	"title": "ML Sklearn Predict",
+	"protocol": "preprocessor_ml_sklearn://",
+	"classType": ["preprocessor"],
+	"capabilities": [],
+	"register": "filter",
+	"node": "python",
+	"path": "nodes.ml_sklearn",
+	"prefix": "Preprocessor",
+	"description": ["Applies a trained scikit-learn model to input text. ",
+		"Loads a pickled model from a configurable path and runs inference, ",
+		"returning the prediction as output. Falls back to original input ",
+		"if no model is found or inference fails."],
+	"icon": "preprocessor-code.svg",
+	"documentation": "https://docs.rocketride.org",
+	"tile": [],
+	"lanes": {
+		"answers": ["answers"]
+	},
+	"input": [
+		{
+			"lane": "answers",
+			"min": 0,
+			"output": [
+				{
+					"lane": "answers"
+				}
+			]
+		}
+	],
+	"preconfig": {
+		"default": "default",
+		"profiles": {
+			"default": {
+				"model_path": ""
+			}
+		}
+	},
+	"test": {
+		"profiles": ["default"],
+		"cases": [
+			{
+				"answers": {
+					"answers": [{"text": "hello world"}]
+				},
+				"expect": {
+					"answers": {
+						"notEmpty": true
+					}
+				}
+			}
+		]
+	},
+	"fields": {
+		"ml_sklearn.model_path": {
+			"type": "string",
+			"title": "Model file path (.pkl)",
+			"default": ""
+		},
+		"ml_sklearn.default": {
+			"object": "default",
+			"properties": ["ml_sklearn.model_path"]
+		}
+	},
+	"shape": [
+		{
+			"section": "Pipe",
+			"title": "ML Sklearn",
+			"properties": ["ml_sklearn.model_path"]
+		}
+	]
+}
\ No newline at end of file
diff --git a/nodes/src/nodes/ml_sklearn/test_ml_sklearn.py b/nodes/src/nodes/ml_sklearn/test_ml_sklearn.py
new file mode 100644
index 000000000..e988514f8
--- /dev/null
+++ b/nodes/src/nodes/ml_sklearn/test_ml_sklearn.py
@@ -0,0 +1,76 @@
+# =============================================================================
+# Standalone test for ml_sklearn — no pytest needed
+# Run: python nodes/src/nodes/ml_sklearn/test_ml_sklearn.py
+# =============================================================================
+import os
+import sys
+import joblib
+import numpy as np
+from sklearn.linear_model import LinearRegression
+
+# ---- Build a temp model ----
+import tempfile
+
+tmp = tempfile.mkdtemp()
+model_path = os.path.join(tmp, 'model.pkl')
+
+X = np.array([[1.0], [2.0], [3.0], [4.0], [5.0]])
+y = np.array([3.0, 5.0, 7.0, 9.0, 11.0])
+model = LinearRegression()
+model.fit(X, y)
+joblib.dump(model, model_path)
+
+# ---- Import PreProcessor directly ----
+import importlib.util
+
+spec = importlib.util.spec_from_file_location(
+    'mlcode',  # 'mlcode' use karo — 'code' nahi, clash hoga
+    os.path.join(os.path.dirname(__file__), 'code.py') if '__file__' in dir() else 'nodes/src/nodes/ml_sklearn/code.py',
+)
+mod = importlib.util.module_from_spec(spec)
+spec.loader.exec_module(mod)
+PreProcessor = mod.PreProcessor
+
+# ---- Tests ----
+passed = 0
+failed = 0
+
+
+def check(name, condition):
+    global passed, failed
+    if condition:
+        print(f'  PASS  {name}')
+        passed += 1
+    else:
+        print(f'  FAIL  {name}')
+        failed += 1
+
+
+print('\nRunning ml_sklearn tests...\n')
+
+# Test 1: Normal prediction
+pp = PreProcessor({'model_path': model_path})
+result = pp.process('4.0')
+check('predict(4.0) ≈ 9.0', abs(float(result) - 9.0) < 0.1)
+
+# Test 2: predict(6.0)
+result2 = pp.process('6.0')
+check('predict(6.0) ≈ 13.0', abs(float(result2) - 13.0) < 0.1)
+
+# Test 3: No model fallback
+pp2 = PreProcessor({'model_path': ''})
+result3 = pp2.process('hello world')
+check('no model → returns input unchanged', result3 == 'hello world')
+
+# Test 4: Bad input fallback
+pp3 = PreProcessor({'model_path': model_path})
+result4 = pp3.process('not a number')
+check('bad input → returns input unchanged', result4 == 'not a number')
+
+# ---- Summary ----
+print(f'\n{passed} passed, {failed} failed')
+if failed == 0:
+    print('All tests passed! ✅')
+else:
+    print('Some tests failed ❌')
+    sys.exit(1)