diff --git a/pretab/transformers/binning/binning.py b/pretab/transformers/binning/binning.py
index d68d52b..9f1fbf7 100644
--- a/pretab/transformers/binning/binning.py
+++ b/pretab/transformers/binning/binning.py
@@ -3,6 +3,7 @@
 import numpy as np
 import pandas as pd
 from sklearn.base import BaseEstimator, TransformerMixin
+from sklearn.utils.validation import check_is_fitted
 
 from ...core.exceptions import InsufficientSamplesError, InvalidParamError, PretabDataError
 from ...core.params import UNSET, AliasResolverMixin
@@ -144,14 +145,17 @@ def get_feature_names_out(self, input_features=None):
 
         Parameters
         ----------
-        input_features : list of str
-            The names of the input features.
+        input_features : list of str or None
+            The names of the input features. When ``None``, names of the form
+            ``x0, x1, ...`` are generated, matching the rest of the package and
+            scikit-learn's contract that the call works with no arguments.
 
         Returns
         -------
-        input_features : ndarray of shape (n_features,)
+        feature_names : ndarray of shape (n_features,)
             The names of the output features after transformation.
         """
+        check_is_fitted(self, "n_features_in_")
         if input_features is None:
-            raise InvalidParamError("input_features must be specified")
-        return input_features
+            input_features = [f"x{i}" for i in range(self.n_features_in_)]
+        return np.asarray(input_features, dtype=object)
diff --git a/tests/test_custombin_transformer.py b/tests/test_custombin_transformer.py
index 1be1ec5..30d19a2 100644
--- a/tests/test_custombin_transformer.py
+++ b/tests/test_custombin_transformer.py
@@ -69,13 +69,47 @@ def test_custom_bin_transformer_feature_names_out():
     transformer = CustomBinTransformer(output_dim=3)
     transformer.fit(np.array([[0.2]]))
     names = transformer.get_feature_names_out(["feature1"])
-    assert names == ["feature1"]
+    assert isinstance(names, np.ndarray)
+    assert list(names) == ["feature1"]
 
 
-def test_custom_bin_transformer_feature_names_out_raises():
-    transformer = CustomBinTransformer(output_dim=3)
-    with pytest.raises(ValueError):
-        transformer.get_feature_names_out()
+def test_custom_bin_transformer_feature_names_out_before_fit_raises():
+    from sklearn.exceptions import NotFittedError
+
+    with pytest.raises(NotFittedError):
+        CustomBinTransformer(output_dim=3).get_feature_names_out()
+
+
+# --------------------------------------------------------------------------- #
+# ``get_feature_names_out()`` must work with no arguments and return an ndarray.
+#
+# It used to raise without ``input_features`` and return a plain list with them,
+# which broke ``Pipeline.get_feature_names_out()``. The same fix was already
+# applied to NoTransformer / ToFloatTransformer / ContinuousOrdinalTransformer.
+# --------------------------------------------------------------------------- #
+def test_feature_names_out_defaults_to_generated_names():
+    transformer = CustomBinTransformer(output_dim=3).fit(np.linspace(0, 1, 10).reshape(-1, 1))
+
+    names = transformer.get_feature_names_out()
+
+    assert isinstance(names, np.ndarray)
+    assert list(names) == ["x0"]
+
+
+def test_feature_names_out_works_inside_a_pipeline():
+    from sklearn.pipeline import Pipeline
+
+    X = np.linspace(0, 1, 20).reshape(-1, 1)
+    pipe = Pipeline([("bin", CustomBinTransformer(output_dim=4))]).fit(X)
+
+    assert list(np.asarray(pipe.get_feature_names_out())) == ["x0"]
+
+
+def test_feature_names_out_length_matches_transform_width():
+    transformer = CustomBinTransformer(output_dim=4).fit(np.linspace(0, 1, 20).reshape(-1, 1))
+
+    width = transformer.transform(np.linspace(0, 1, 20).reshape(-1, 1)).shape[1]
+    assert len(transformer.get_feature_names_out()) == width
 
 
 def test_custom_bin_transformer_is_sklearn_compatible():