diff --git a/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/index.rst b/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/index.rst
index 6300f77d686..f6caabe324d 100644
--- a/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/index.rst
+++ b/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/index.rst
@@ -7,3 +7,4 @@ nvtext
     edit_distance
     generate_ngrams
     jaccard
+    minhash
diff --git a/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/minhash.rst b/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/minhash.rst
new file mode 100644
index 00000000000..b8ec02fca35
--- /dev/null
+++ b/docs/cudf/source/user_guide/api_docs/pylibcudf/nvtext/minhash.rst
@@ -0,0 +1,6 @@
+=======
+minhash
+=======
+
+.. automodule:: pylibcudf.nvtext.minhash
+   :members:
diff --git a/python/cudf/cudf/_lib/nvtext/minhash.pyx b/python/cudf/cudf/_lib/nvtext/minhash.pyx
index 59cb8d51440..5e39cafa47b 100644
--- a/python/cudf/cudf/_lib/nvtext/minhash.pyx
+++ b/python/cudf/cudf/_lib/nvtext/minhash.pyx
@@ -2,93 +2,44 @@
 
 from cudf.core.buffer import acquire_spill_lock
 
-from libcpp.memory cimport unique_ptr
-from libcpp.utility cimport move
-
-from pylibcudf.libcudf.column.column cimport column
-from pylibcudf.libcudf.column.column_view cimport column_view
-from pylibcudf.libcudf.nvtext.minhash cimport (
-    minhash as cpp_minhash,
-    minhash64 as cpp_minhash64,
-    word_minhash as cpp_word_minhash,
-    word_minhash64 as cpp_word_minhash64,
-)
-from pylibcudf.libcudf.types cimport size_type
-
 from cudf._lib.column cimport Column
 
-
-@acquire_spill_lock()
-def minhash(Column strings, Column seeds, int width):
-
-    cdef column_view c_strings = strings.view()
-    cdef size_type c_width = width
-    cdef column_view c_seeds = seeds.view()
-    cdef unique_ptr[column] c_result
-
-    with nogil:
-        c_result = move(
-            cpp_minhash(
-                c_strings,
-                c_seeds,
-                c_width
-            )
-        )
-
-    return Column.from_unique_ptr(move(c_result))
+from pylibcudf import nvtext
 
 
 @acquire_spill_lock()
-def minhash64(Column strings, Column seeds, int width):
-
-    cdef column_view c_strings = strings.view()
-    cdef size_type c_width = width
-    cdef column_view c_seeds = seeds.view()
-    cdef unique_ptr[column] c_result
+def minhash(Column input, Column seeds, int width=4):
+    result = nvtext.minhash.minhash(
+        input.to_pylibcudf(mode="read"),
+        seeds.to_pylibcudf(mode="read"),
+        width,
+    )
+    return Column.from_pylibcudf(result)
 
-    with nogil:
-        c_result = move(
-            cpp_minhash64(
-                c_strings,
-                c_seeds,
-                c_width
-            )
-        )
 
-    return Column.from_unique_ptr(move(c_result))
+@acquire_spill_lock()
+def minhash64(Column input, Column seeds, int width=4):
+    result = nvtext.minhash.minhash64(
+        input.to_pylibcudf(mode="read"),
+        seeds.to_pylibcudf(mode="read"),
+        width,
+    )
+    return Column.from_pylibcudf(result)
 
 
 @acquire_spill_lock()
 def word_minhash(Column input, Column seeds):
-
-    cdef column_view c_input = input.view()
-    cdef column_view c_seeds = seeds.view()
-    cdef unique_ptr[column] c_result
-
-    with nogil:
-        c_result = move(
-            cpp_word_minhash(
-                c_input,
-                c_seeds
-            )
-        )
-
-    return Column.from_unique_ptr(move(c_result))
+    result = nvtext.minhash.word_minhash(
+        input.to_pylibcudf(mode="read"),
+        seeds.to_pylibcudf(mode="read"),
+    )
+    return Column.from_pylibcudf(result)
 
 
 @acquire_spill_lock()
 def word_minhash64(Column input, Column seeds):
-
-    cdef column_view c_input = input.view()
-    cdef column_view c_seeds = seeds.view()
-    cdef unique_ptr[column] c_result
-
-    with nogil:
-        c_result = move(
-            cpp_word_minhash64(
-                c_input,
-                c_seeds
-            )
-        )
-
-    return Column.from_unique_ptr(move(c_result))
+    result = nvtext.minhash.word_minhash64(
+        input.to_pylibcudf(mode="read"),
+        seeds.to_pylibcudf(mode="read"),
+    )
+    return Column.from_pylibcudf(result)
diff --git a/python/pylibcudf/pylibcudf/libcudf/concatenate.pxd b/python/pylibcudf/pylibcudf/libcudf/concatenate.pxd
index a09b6c01392..def292148c5 100644
--- a/python/pylibcudf/pylibcudf/libcudf/concatenate.pxd
+++ b/python/pylibcudf/pylibcudf/libcudf/concatenate.pxd
@@ -4,7 +4,7 @@ from libcpp.memory cimport unique_ptr
 from libcpp.vector cimport vector
 from pylibcudf.libcudf.column.column cimport column, column_view
 from pylibcudf.libcudf.table.table cimport table, table_view
-from pylibcudf.libcudf.utilities.host_span cimport host_span
+from pylibcudf.libcudf.utilities.span cimport host_span
 
 from rmm.librmm.device_buffer cimport device_buffer
 
diff --git a/python/pylibcudf/pylibcudf/libcudf/groupby.pxd b/python/pylibcudf/pylibcudf/libcudf/groupby.pxd
index 848462131fe..17ea33a2066 100644
--- a/python/pylibcudf/pylibcudf/libcudf/groupby.pxd
+++ b/python/pylibcudf/pylibcudf/libcudf/groupby.pxd
@@ -22,7 +22,6 @@ from pylibcudf.libcudf.types cimport (
     size_type,
     sorted,
 )
-from pylibcudf.libcudf.utilities.host_span cimport host_span
 
 # workaround for https://github.com/cython/cython/issues/3885
 ctypedef const scalar constscalar
diff --git a/python/pylibcudf/pylibcudf/libcudf/nvtext/minhash.pxd b/python/pylibcudf/pylibcudf/libcudf/nvtext/minhash.pxd
index f2dd22f43aa..41250037dcf 100644
--- a/python/pylibcudf/pylibcudf/libcudf/nvtext/minhash.pxd
+++ b/python/pylibcudf/pylibcudf/libcudf/nvtext/minhash.pxd
@@ -1,13 +1,21 @@
 # Copyright (c) 2023-2024, NVIDIA CORPORATION.
 
+from libc.stdint cimport uint32_t, uint64_t
 from libcpp.memory cimport unique_ptr
 from pylibcudf.libcudf.column.column cimport column
 from pylibcudf.libcudf.column.column_view cimport column_view
+from pylibcudf.libcudf.scalar.scalar cimport numeric_scalar
 from pylibcudf.libcudf.types cimport size_type
 
 
 cdef extern from "nvtext/minhash.hpp" namespace "nvtext" nogil:
 
+    cdef unique_ptr[column] minhash(
+        const column_view &strings,
+        const numeric_scalar[uint32_t] seed,
+        const size_type width,
+    ) except +
+
     cdef unique_ptr[column] minhash(
         const column_view &strings,
         const column_view &seeds,
@@ -20,6 +28,12 @@ cdef extern from "nvtext/minhash.hpp" namespace "nvtext" nogil:
         const size_type width,
     ) except +
 
+    cdef unique_ptr[column] minhash64(
+        const column_view &strings,
+        const numeric_scalar[uint64_t] seed,
+        const size_type width,
+    ) except +
+
     cdef unique_ptr[column] word_minhash(
         const column_view &input,
         const column_view &seeds
diff --git a/python/pylibcudf/pylibcudf/libcudf/utilities/host_span.pxd b/python/pylibcudf/pylibcudf/libcudf/utilities/span.pxd
similarity index 100%
rename from python/pylibcudf/pylibcudf/libcudf/utilities/host_span.pxd
rename to python/pylibcudf/pylibcudf/libcudf/utilities/span.pxd
diff --git a/python/pylibcudf/pylibcudf/nvtext/CMakeLists.txt b/python/pylibcudf/pylibcudf/nvtext/CMakeLists.txt
index 9913e1fbadb..7fd65beeeb0 100644
--- a/python/pylibcudf/pylibcudf/nvtext/CMakeLists.txt
+++ b/python/pylibcudf/pylibcudf/nvtext/CMakeLists.txt
@@ -12,7 +12,7 @@
 # the License.
 # =============================================================================
 
-set(cython_sources edit_distance.pyx generate_ngrams.pyx jaccard.pyx)
+set(cython_sources edit_distance.pyx generate_ngrams.pyx jaccard.pyx minhash.pyx)
 
 set(linked_libraries cudf::cudf)
 rapids_cython_create_modules(
diff --git a/python/pylibcudf/pylibcudf/nvtext/__init__.pxd b/python/pylibcudf/pylibcudf/nvtext/__init__.pxd
index 5f1762b1e3d..9eed1da1ab5 100644
--- a/python/pylibcudf/pylibcudf/nvtext/__init__.pxd
+++ b/python/pylibcudf/pylibcudf/nvtext/__init__.pxd
@@ -1,9 +1,10 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
 
-from . cimport edit_distance, generate_ngrams, jaccard
+from . cimport edit_distance, generate_ngrams, jaccard, minhash
 
 __all__ = [
     "edit_distance",
     "generate_ngrams",
     "jaccard",
+    "minhash"
 ]
diff --git a/python/pylibcudf/pylibcudf/nvtext/__init__.py b/python/pylibcudf/pylibcudf/nvtext/__init__.py
index 1c0ddb1e5a4..a3a2363f7ef 100644
--- a/python/pylibcudf/pylibcudf/nvtext/__init__.py
+++ b/python/pylibcudf/pylibcudf/nvtext/__init__.py
@@ -1,9 +1,10 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
 
-from . import edit_distance, generate_ngrams, jaccard
+from . import edit_distance, generate_ngrams, jaccard, minhash
 
 __all__ = [
     "edit_distance",
     "generate_ngrams",
     "jaccard",
+    "minhash",
 ]
diff --git a/python/pylibcudf/pylibcudf/nvtext/minhash.pxd b/python/pylibcudf/pylibcudf/nvtext/minhash.pxd
new file mode 100644
index 00000000000..97e8c9dc83c
--- /dev/null
+++ b/python/pylibcudf/pylibcudf/nvtext/minhash.pxd
@@ -0,0 +1,18 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from libc.stdint cimport uint32_t, uint64_t
+from pylibcudf.column cimport Column
+from pylibcudf.libcudf.types cimport size_type
+from pylibcudf.scalar cimport Scalar
+
+ctypedef fused ColumnOrScalar:
+    Column
+    Scalar
+
+cpdef Column minhash(Column input, ColumnOrScalar seeds, size_type width=*)
+
+cpdef Column minhash64(Column input, ColumnOrScalar seeds, size_type width=*)
+
+cpdef Column word_minhash(Column input, Column seeds)
+
+cpdef Column word_minhash64(Column input, Column seeds)
diff --git a/python/pylibcudf/pylibcudf/nvtext/minhash.pyx b/python/pylibcudf/pylibcudf/nvtext/minhash.pyx
new file mode 100644
index 00000000000..5fabf6a3f89
--- /dev/null
+++ b/python/pylibcudf/pylibcudf/nvtext/minhash.pyx
@@ -0,0 +1,160 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from libc.stdint cimport uint32_t, uint64_t
+from libcpp.memory cimport unique_ptr
+from libcpp.utility cimport move
+from pylibcudf.column cimport Column
+from pylibcudf.libcudf.column.column cimport column
+from pylibcudf.libcudf.nvtext.minhash cimport (
+    minhash as cpp_minhash,
+    minhash64 as cpp_minhash64,
+    word_minhash as cpp_word_minhash,
+    word_minhash64 as cpp_word_minhash64,
+)
+from pylibcudf.libcudf.scalar.scalar cimport numeric_scalar
+from pylibcudf.libcudf.types cimport size_type
+from pylibcudf.scalar cimport Scalar
+
+from cython.operator import dereference
+
+
+cpdef Column minhash(Column input, ColumnOrScalar seeds, size_type width=4):
+    """
+    Returns the minhash values for each string per seed.
+    This function uses MurmurHash3_x86_32 for the hash algorithm.
+
+    For details, see :cpp:func:`minhash`.
+
+    Parameters
+    ----------
+    input : Column
+        Strings column to compute minhash
+    seeds : Column or Scalar
+        Seed value(s) used for the hash algorithm.
+    width : size_type
+        Character width used for apply substrings;
+        Default is 4 characters.
+
+    Returns
+    -------
+    Column
+        List column of minhash values for each string per seed
+    """
+    cdef unique_ptr[column] c_result
+
+    if not isinstance(seeds, (Column, Scalar)):
+        raise TypeError("Must pass a Column or Scalar")
+
+    with nogil:
+        c_result = move(
+            cpp_minhash(
+                input.view(),
+                seeds.view() if ColumnOrScalar is Column else
+                dereference(<numeric_scalar[uint32_t]*>seeds.c_obj.get()),
+                width
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+cpdef Column minhash64(Column input, ColumnOrScalar seeds, size_type width=4):
+    """
+    Returns the minhash values for each string per seed.
+    This function uses MurmurHash3_x64_128 for the hash algorithm.
+
+    For details, see :cpp:func:`minhash64`.
+
+    Parameters
+    ----------
+    input : Column
+        Strings column to compute minhash
+    seeds : Column or Scalar
+        Seed value(s) used for the hash algorithm.
+    width : size_type
+        Character width used for apply substrings;
+        Default is 4 characters.
+
+    Returns
+    -------
+    Column
+        List column of minhash values for each string per seed
+    """
+    cdef unique_ptr[column] c_result
+
+    if not isinstance(seeds, (Column, Scalar)):
+        raise TypeError("Must pass a Column or Scalar")
+
+    with nogil:
+        c_result = move(
+            cpp_minhash64(
+                input.view(),
+                seeds.view() if ColumnOrScalar is Column else
+                dereference(<numeric_scalar[uint64_t]*>seeds.c_obj.get()),
+                width
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+cpdef Column word_minhash(Column input, Column seeds):
+    """
+    Returns the minhash values for each row of strings per seed.
+    This function uses MurmurHash3_x86_32 for the hash algorithm.
+
+    For details, see :cpp:func:`word_minhash`.
+
+    Parameters
+    ----------
+    input : Column
+        Lists column of strings to compute minhash
+    seeds : Column or Scalar
+        Seed values used for the hash algorithm.
+
+    Returns
+    -------
+    Column
+        List column of minhash values for each string per seed
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(
+            cpp_word_minhash(
+                input.view(),
+                seeds.view()
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
+
+cpdef Column word_minhash64(Column input, Column seeds):
+    """
+    Returns the minhash values for each row of strings per seed.
+    This function uses MurmurHash3_x64_128 for the hash algorithm though
+    only the first 64-bits of the hash are used in computing the output.
+
+    For details, see :cpp:func:`word_minhash64`.
+
+    Parameters
+    ----------
+    input : Column
+        Lists column of strings to compute minhash
+    seeds : Column or Scalar
+        Seed values used for the hash algorithm.
+
+    Returns
+    -------
+    Column
+        List column of minhash values for each string per seed
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(
+            cpp_word_minhash64(
+                input.view(),
+                seeds.view()
+            )
+        )
+
+    return Column.from_libcudf(move(c_result))
diff --git a/python/pylibcudf/pylibcudf/tests/test_nvtext_minhash.py b/python/pylibcudf/pylibcudf/tests/test_nvtext_minhash.py
new file mode 100644
index 00000000000..4e389a63f90
--- /dev/null
+++ b/python/pylibcudf/pylibcudf/tests/test_nvtext_minhash.py
@@ -0,0 +1,54 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+import pyarrow as pa
+import pylibcudf as plc
+import pytest
+
+
+@pytest.fixture(scope="module", params=[pa.uint32(), pa.uint64()])
+def minhash_input_data(request):
+    input_arr = pa.array(["foo", "bar", "foo foo", "bar bar"])
+    seeds = pa.array([2, 3, 4, 5], request.param)
+    return input_arr, seeds, request.param
+
+
+@pytest.fixture(scope="module", params=[pa.uint32(), pa.uint64()])
+def word_minhash_input_data(request):
+    input_arr = pa.array([["foo", "bar"], ["foo foo", "bar bar"]])
+    seeds = pa.array([2, 3, 4, 5], request.param)
+    return input_arr, seeds, request.param
+
+
+@pytest.mark.parametrize("width", [5, 12])
+def test_minhash(minhash_input_data, width):
+    input_arr, seeds, seed_type = minhash_input_data
+    minhash_func = (
+        plc.nvtext.minhash.minhash
+        if seed_type == pa.uint32()
+        else plc.nvtext.minhash.minhash64
+    )
+    result = minhash_func(
+        plc.interop.from_arrow(input_arr), plc.interop.from_arrow(seeds), width
+    )
+    pa_result = plc.interop.to_arrow(result)
+    assert all(len(got) == len(seeds) for got, s in zip(pa_result, input_arr))
+    assert pa_result.type == pa.list_(
+        pa.field("element", seed_type, nullable=False)
+    )
+
+
+def test_word_minhash(word_minhash_input_data):
+    input_arr, seeds, seed_type = word_minhash_input_data
+    word_minhash_func = (
+        plc.nvtext.minhash.word_minhash
+        if seed_type == pa.uint32()
+        else plc.nvtext.minhash.word_minhash64
+    )
+    result = word_minhash_func(
+        plc.interop.from_arrow(input_arr), plc.interop.from_arrow(seeds)
+    )
+    pa_result = plc.interop.to_arrow(result)
+    assert all(len(got) == len(seeds) for got, s in zip(pa_result, input_arr))
+    assert pa_result.type == pa.list_(
+        pa.field("element", seed_type, nullable=False)
+    )