diff --git a/stubs/seaborn/seaborn/_stats/density.pyi b/stubs/seaborn/seaborn/_stats/density.pyi
index 7e8d12e30..386424ccb 100644
--- a/stubs/seaborn/seaborn/_stats/density.pyi
+++ b/stubs/seaborn/seaborn/_stats/density.pyi
@@ -1,13 +1,12 @@
-from collections.abc import Callable
 from dataclasses import dataclass
 
 from seaborn._stats.base import Stat
-from seaborn.external.kde import gaussian_kde
+from seaborn.external.kde import _BwMethodType
 
 @dataclass
 class KDE(Stat):
     bw_adjust: float = 1
-    bw_method: str | float | Callable[[gaussian_kde], float] = "scott"
+    bw_method: _BwMethodType = "scott"
     common_norm: bool | list[str] = True
     common_grid: bool | list[str] = True
     gridsize: int | None = 200
diff --git a/stubs/seaborn/seaborn/distributions.pyi b/stubs/seaborn/seaborn/distributions.pyi
index 78c20b4d5..0f849f2c1 100644
--- a/stubs/seaborn/seaborn/distributions.pyi
+++ b/stubs/seaborn/seaborn/distributions.pyi
@@ -1,28 +1,37 @@
-from _typeshed import Incomplete
 from collections.abc import Iterable
-from typing import Any
-from typing_extensions import Literal, deprecated
+from typing import Any, Protocol, TypeVar
+from typing_extensions import Literal, TypeAlias, deprecated
 
 from matplotlib.axes import Axes
-from matplotlib.colors import Colormap
+from matplotlib.colors import Colormap, Normalize
 from matplotlib.typing import ColorType
+from numpy.typing import ArrayLike
 
+from ._core.typing import ColumnName, DataSource
 from .axisgrid import FacetGrid
-from .utils import _LogScale, _Palette
+from .external.kde import _BwMethodType
+from .utils import _DataSourceWideForm, _LogScale, _Palette, _Vector
 
 __all__ = ["displot", "histplot", "kdeplot", "ecdfplot", "rugplot", "distplot"]
 
+_T = TypeVar("_T")
+_OneOrPair: TypeAlias = _T | tuple[_T, _T]
+
+class _Fit(Protocol):
+    def fit(self, a: ArrayLike) -> tuple[ArrayLike, ...]: ...
+    def pdf(self, x: ArrayLike, *params: ArrayLike) -> ArrayLike: ...
+
 def histplot(
-    data: Incomplete | None = None,
+    data: DataSource | _DataSourceWideForm | None = None,
     *,
-    x: Incomplete | None = None,
-    y: Incomplete | None = None,
-    hue: Incomplete | None = None,
-    weights: Incomplete | None = None,
+    x: ColumnName | _Vector | None = None,
+    y: ColumnName | _Vector | None = None,
+    hue: ColumnName | _Vector | None = None,
+    weights: ColumnName | _Vector | None = None,
     stat: str = "count",
-    bins: Incomplete = "auto",
+    bins: _OneOrPair[str | int | ArrayLike] = "auto",
     binwidth: float | tuple[float, float] | None = None,
-    binrange: Incomplete | None = None,
+    binrange: _OneOrPair[tuple[float, float]] | None = None,
     discrete: bool | None = None,
     cumulative: bool = False,
     common_bins: bool = True,
@@ -41,8 +50,8 @@ def histplot(
     cbar_ax: Axes | None = None,
     cbar_kws: dict[str, Any] | None = None,
     palette: _Palette | Colormap | None = None,
-    hue_order: Iterable[str] | None = None,
-    hue_norm: Incomplete | None = None,
+    hue_order: Iterable[ColumnName] | None = None,
+    hue_norm: tuple[float, float] | Normalize | None = None,
     color: ColorType | None = None,
     log_scale: _LogScale | None = None,
     legend: bool = True,
@@ -50,30 +59,30 @@ def histplot(
     **kwargs: Any,
 ) -> Axes: ...
 def kdeplot(
-    data: Incomplete | None = None,
+    data: DataSource | _DataSourceWideForm | None = None,
     *,
-    x: Incomplete | None = None,
-    y: Incomplete | None = None,
-    hue: Incomplete | None = None,
-    weights: Incomplete | None = None,
+    x: ColumnName | _Vector | None = None,
+    y: ColumnName | _Vector | None = None,
+    hue: ColumnName | _Vector | None = None,
+    weights: ColumnName | _Vector | None = None,
     palette: _Palette | Colormap | None = None,
-    hue_order: Iterable[str] | None = None,
-    hue_norm: Incomplete | None = None,
+    hue_order: Iterable[ColumnName] | None = None,
+    hue_norm: tuple[float, float] | Normalize | None = None,
     color: ColorType | None = None,
     fill: bool | None = None,
     multiple: Literal["layer", "stack", "fill"] = "layer",
     common_norm: bool = True,
     common_grid: bool = False,
     cumulative: bool = False,
-    bw_method: str = "scott",
+    bw_method: _BwMethodType = "scott",
     bw_adjust: float = 1,
     warn_singular: bool = True,
     log_scale: _LogScale | None = None,
-    levels: int | Iterable[int] = 10,
+    levels: int | Iterable[float] = 10,
     thresh: float = 0.05,
     gridsize: int = 200,
     cut: float = 3,
-    clip: Incomplete | None = None,
+    clip: _OneOrPair[tuple[float | None, float | None]] | None = None,
     legend: bool = True,
     cbar: bool = False,
     cbar_ax: Axes | None = None,
@@ -82,58 +91,58 @@ def kdeplot(
     **kwargs: Any,
 ) -> Axes: ...
 def ecdfplot(
-    data: Incomplete | None = None,
+    data: DataSource | _DataSourceWideForm | None = None,
     *,
-    x: Incomplete | None = None,
-    y: Incomplete | None = None,
-    hue: Incomplete | None = None,
-    weights: Incomplete | None = None,
-    stat: Literal["proportion", "count"] = "proportion",
+    x: ColumnName | _Vector | None = None,
+    y: ColumnName | _Vector | None = None,
+    hue: ColumnName | _Vector | None = None,
+    weights: ColumnName | _Vector | None = None,
+    stat: Literal["proportion", "percent", "count"] = "proportion",
     complementary: bool = False,
     palette: _Palette | Colormap | None = None,
-    hue_order: Iterable[str] | None = None,
-    hue_norm: Incomplete | None = None,
+    hue_order: Iterable[ColumnName] | None = None,
+    hue_norm: tuple[float, float] | Normalize | None = None,
     log_scale: _LogScale | None = None,
     legend: bool = True,
     ax: Axes | None = None,
     **kwargs: Any,
 ) -> Axes: ...
 def rugplot(
-    data: Incomplete | None = None,
+    data: DataSource | _DataSourceWideForm | None = None,
     *,
-    x: Incomplete | None = None,
-    y: Incomplete | None = None,
-    hue: Incomplete | None = None,
+    x: ColumnName | _Vector | None = None,
+    y: ColumnName | _Vector | None = None,
+    hue: ColumnName | _Vector | None = None,
     height: float = 0.025,
     expand_margins: bool = True,
     palette: _Palette | Colormap | None = None,
-    hue_order: Iterable[str] | None = None,
-    hue_norm: Incomplete | None = None,
+    hue_order: Iterable[ColumnName] | None = None,
+    hue_norm: tuple[float, float] | Normalize | None = None,
     legend: bool = True,
     ax: Axes | None = None,
     **kwargs: Any,
 ) -> Axes: ...
 def displot(
-    data: Incomplete | None = None,
+    data: DataSource | _DataSourceWideForm | None = None,
     *,
-    x: Incomplete | None = None,
-    y: Incomplete | None = None,
-    hue: Incomplete | None = None,
-    row: Incomplete | None = None,
-    col: Incomplete | None = None,
-    weights: Incomplete | None = None,
+    x: ColumnName | _Vector | None = None,
+    y: ColumnName | _Vector | None = None,
+    hue: ColumnName | _Vector | None = None,
+    row: ColumnName | _Vector | None = None,
+    col: ColumnName | _Vector | None = None,
+    weights: ColumnName | _Vector | None = None,
     kind: Literal["hist", "kde", "ecdf"] = "hist",
     rug: bool = False,
     rug_kws: dict[str, Any] | None = None,
     log_scale: _LogScale | None = None,
     legend: bool = True,
     palette: _Palette | Colormap | None = None,
-    hue_order: Iterable[str] | None = None,
-    hue_norm: Incomplete | None = None,
+    hue_order: Iterable[ColumnName] | None = None,
+    hue_norm: tuple[float, float] | Normalize | None = None,
     color: ColorType | None = None,
     col_wrap: int | None = None,
-    row_order: Iterable[str] | None = None,
-    col_order: Iterable[str] | None = None,
+    row_order: Iterable[ColumnName] | None = None,
+    col_order: Iterable[ColumnName] | None = None,
     height: float = 5,
     aspect: float = 1,
     facet_kws: dict[str, Any] | None = None,
@@ -141,12 +150,12 @@ def displot(
 ) -> FacetGrid: ...
 @deprecated("Function `distplot` is deprecated and will be removed in seaborn v0.14.0")
 def distplot(
-    a: Incomplete | None = None,
-    bins: Incomplete | None = None,
+    a: ArrayLike | None = None,
+    bins: ArrayLike | None = None,
     hist: bool = True,
     kde: bool = True,
     rug: bool = False,
-    fit: Incomplete | None = None,
+    fit: _Fit | None = None,
     hist_kws: dict[str, Any] | None = None,
     kde_kws: dict[str, Any] | None = None,
     rug_kws: dict[str, Any] | None = None,
@@ -155,7 +164,7 @@ def distplot(
     vertical: bool = False,
     norm_hist: bool = False,
     axlabel: str | Literal[False] | None = None,
-    label: Incomplete | None = None,
+    label: str | None = None,
     ax: Axes | None = None,
-    x: Incomplete | None = None,
+    x: ArrayLike | None = None,
 ) -> Axes: ...
diff --git a/stubs/seaborn/seaborn/external/kde.pyi b/stubs/seaborn/seaborn/external/kde.pyi
index 92d622779..81b757304 100644
--- a/stubs/seaborn/seaborn/external/kde.pyi
+++ b/stubs/seaborn/seaborn/external/kde.pyi
@@ -1,4 +1,5 @@
 from collections.abc import Callable
+from typing import Any, Protocol
 from typing_extensions import Literal, TypeAlias
 
 import numpy as np
@@ -6,8 +7,25 @@ from numpy.typing import ArrayLike, NDArray
 
 __all__ = ["gaussian_kde"]
 
-_Scalar: TypeAlias = np.generic | bool | int | float | complex | bytes | memoryview  # see np.isscalar
-_BwMethodType: TypeAlias = Literal["scott", "silverman"] | Callable[[gaussian_kde], object] | _Scalar | None
+# define a "Gaussian KDE" protocol so that we can also pass `scipy.stats.gaussian_kde` to
+# functions that expect it without adding a dependency on scipy
+class _GaussianKDELike(Protocol):
+    dataset: NDArray[np.float64]
+    def __init__(self, dataset: ArrayLike, bw_method: Any | None = ..., weights: ArrayLike | None = ...) -> None: ...
+    def evaluate(self, points: ArrayLike) -> NDArray[Any]: ...
+    def __call__(self, points: ArrayLike) -> NDArray[Any]: ...
+    def scotts_factor(self) -> float: ...
+    def silverman_factor(self) -> float: ...
+    def covariance_factor(self) -> float: ...
+    def pdf(self, x: ArrayLike) -> NDArray[Any]: ...
+    def set_bandwidth(self, bw_method: Any | None = ...) -> None: ...
+    @property
+    def weights(self) -> NDArray[Any]: ...
+    @property
+    def neff(self) -> NDArray[Any]: ...
+
+_Scalar: TypeAlias = float | np.number[Any]
+_BwMethodType: TypeAlias = Literal["scott", "silverman"] | Callable[[_GaussianKDELike], _Scalar] | _Scalar | None
 
 class gaussian_kde:
     dataset: NDArray[np.float64]