diff --git a/docs/cudf/source/user_guide/api_docs/pylibcudf/index.rst b/docs/cudf/source/user_guide/api_docs/pylibcudf/index.rst
index 58fea77adaa..b6ad1157511 100644
--- a/docs/cudf/source/user_guide/api_docs/pylibcudf/index.rst
+++ b/docs/cudf/source/user_guide/api_docs/pylibcudf/index.rst
@@ -6,7 +6,7 @@ This page provides API documentation for pylibcudf.
 
 .. toctree::
     :maxdepth: 1
-    :caption: API Documentation
+    :caption: Top-level modules
 
     aggregation
     binaryop
@@ -32,3 +32,9 @@ This page provides API documentation for pylibcudf.
     table
     types
     unary
+
+.. toctree::
+    :maxdepth: 2
+    :caption: Subpackages
+
+    strings/index.rst
diff --git a/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/index.rst b/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/index.rst
new file mode 100644
index 00000000000..8970fc80c0b
--- /dev/null
+++ b/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/index.rst
@@ -0,0 +1,7 @@
+strings
+=======
+
+.. toctree::
+    :maxdepth: 1
+
+    replace
diff --git a/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/replace.rst b/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/replace.rst
new file mode 100644
index 00000000000..9575ec226a7
--- /dev/null
+++ b/docs/cudf/source/user_guide/api_docs/pylibcudf/strings/replace.rst
@@ -0,0 +1,6 @@
+=======
+replace
+=======
+
+.. automodule:: cudf._lib.pylibcudf.strings.replace
+   :members:
diff --git a/python/cudf/cudf/_lib/pylibcudf/strings/CMakeLists.txt b/python/cudf/cudf/_lib/pylibcudf/strings/CMakeLists.txt
index 0e9c1c916f0..c9a983e24f4 100644
--- a/python/cudf/cudf/_lib/pylibcudf/strings/CMakeLists.txt
+++ b/python/cudf/cudf/_lib/pylibcudf/strings/CMakeLists.txt
@@ -12,11 +12,11 @@
 # the License.
 # =============================================================================
 
-set(cython_sources capitalize.pyx case.pyx char_types.pyx find.pyx)
+set(cython_sources capitalize.pyx case.pyx char_types.pyx find.pyx replace.pyx)
 
 set(linked_libraries cudf::cudf)
 rapids_cython_create_modules(
   CXX
   SOURCE_FILES "${cython_sources}"
-  LINKED_LIBRARIES "${linked_libraries}" MODULE_PREFIX pylibcudf_ ASSOCIATED_TARGETS cudf
+  LINKED_LIBRARIES "${linked_libraries}" MODULE_PREFIX pylibcudf_strings_ ASSOCIATED_TARGETS cudf
 )
diff --git a/python/cudf/cudf/_lib/pylibcudf/strings/__init__.pxd b/python/cudf/cudf/_lib/pylibcudf/strings/__init__.pxd
index ec3dbc150b5..7563df8a107 100644
--- a/python/cudf/cudf/_lib/pylibcudf/strings/__init__.pxd
+++ b/python/cudf/cudf/_lib/pylibcudf/strings/__init__.pxd
@@ -1,3 +1,3 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
 
-from . cimport capitalize, case, char_types, find
+from . cimport capitalize, case, char_types, find, replace
diff --git a/python/cudf/cudf/_lib/pylibcudf/strings/__init__.py b/python/cudf/cudf/_lib/pylibcudf/strings/__init__.py
index 3793bda0aa4..cb4f0e38f97 100644
--- a/python/cudf/cudf/_lib/pylibcudf/strings/__init__.py
+++ b/python/cudf/cudf/_lib/pylibcudf/strings/__init__.py
@@ -1,3 +1,3 @@
 # Copyright (c) 2024, NVIDIA CORPORATION.
 
-from . import capitalize, case, char_types, find
+from . import capitalize, case, char_types, find, replace
diff --git a/python/cudf/cudf/_lib/pylibcudf/strings/replace.pxd b/python/cudf/cudf/_lib/pylibcudf/strings/replace.pxd
new file mode 100644
index 00000000000..52e2dc3c738
--- /dev/null
+++ b/python/cudf/cudf/_lib/pylibcudf/strings/replace.pxd
@@ -0,0 +1,25 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from cudf._lib.pylibcudf.column cimport Column
+from cudf._lib.pylibcudf.libcudf.types cimport size_type
+from cudf._lib.pylibcudf.scalar cimport Scalar
+
+
+cpdef Column replace(
+    Column input,
+    Scalar target,
+    Scalar repl,
+    size_type maxrepl = *
+)
+cpdef Column replace_multiple(
+    Column input,
+    Column target,
+    Column repl,
+    size_type maxrepl = *
+)
+cpdef Column replace_slice(
+    Column input,
+    Scalar repl = *,
+    size_type start = *,
+    size_type stop = *
+)
diff --git a/python/cudf/cudf/_lib/pylibcudf/strings/replace.pyx b/python/cudf/cudf/_lib/pylibcudf/strings/replace.pyx
new file mode 100644
index 00000000000..c757150a600
--- /dev/null
+++ b/python/cudf/cudf/_lib/pylibcudf/strings/replace.pyx
@@ -0,0 +1,162 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+from libcpp.memory cimport unique_ptr
+from libcpp.utility cimport move
+
+from cudf._lib.pylibcudf.column cimport Column
+from cudf._lib.pylibcudf.libcudf.column.column cimport column
+from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport string_scalar
+from cudf._lib.pylibcudf.libcudf.scalar.scalar_factories cimport (
+    make_string_scalar as cpp_make_string_scalar,
+)
+from cudf._lib.pylibcudf.libcudf.strings.replace cimport (
+    replace as cpp_replace,
+    replace_multiple as cpp_replace_multiple,
+    replace_slice as cpp_replace_slice,
+)
+from cudf._lib.pylibcudf.libcudf.types cimport size_type
+from cudf._lib.pylibcudf.scalar cimport Scalar
+
+
+cpdef Column replace(
+    Column input,
+    Scalar target,
+    Scalar repl,
+    size_type maxrepl = -1
+):
+    """Replaces target string within each string with the specified replacement string.
+
+    Null string entries will return null output string entries.
+
+    For details, see :cpp:func:`replace`.
+
+    Parameters
+    ----------
+    input : Column
+        The input strings
+    target : Scalar
+        String to search for in each string.
+    repl : Scalar
+        String to replace target with.
+    maxrepl : size_type, default -1
+        Maximum times to replace if target appears multiple times in the input string.
+        Default of -1 specifies to replace all occurrences of target in each string.
+
+    Returns
+    -------
+    pylibcudf.Column
+        New string column with target replaced.
+    """
+    cdef:
+        unique_ptr[column] c_result
+        const string_scalar* target_str
+        const string_scalar* repl_str
+
+    target_str = <string_scalar *>(target.c_obj.get())
+    repl_str = <string_scalar *>(repl.c_obj.get())
+
+    with nogil:
+        c_result = move(cpp_replace(
+            input.view(),
+            target_str[0],
+            repl_str[0],
+            maxrepl,
+        ))
+
+    return Column.from_libcudf(move(c_result))
+
+
+cpdef Column replace_multiple(
+    Column input,
+    Column target,
+    Column repl,
+    size_type maxrepl = -1
+):
+    """Replaces target string within each string with the specified replacement string.
+
+    Null string entries will return null output string entries.
+
+    For details, see :cpp:func:`replace_multiple`.
+
+    Parameters
+    ----------
+    input : Column
+        The input strings
+    target : Column
+        Column containing strings to search for in the input column.
+    repl : Column
+        Column containing strings to replace target with.
+        Each target, when found, will be replaced by the value at the
+        corresponding index in the repl Column.
+
+        Must be of the same length as target.
+
+    Returns
+    -------
+    pylibcudf.Column
+        New string column with target replaced.
+    """
+    cdef unique_ptr[column] c_result
+
+    with nogil:
+        c_result = move(cpp_replace_multiple(
+            input.view(),
+            target.view(),
+            repl.view(),
+        ))
+
+    return Column.from_libcudf(move(c_result))
+
+
+cpdef Column replace_slice(
+    Column input,
+    # TODO: default scalar values
+    # https://github.com/rapidsai/cudf/issues/15505
+    Scalar repl = None,
+    size_type start = 0,
+    size_type stop = -1
+):
+    """Replaces each string in the column with the provided repl string
+    within the [start,stop) character position range.
+
+    Null string entries will return null output string entries.
+    This function can be used to insert a string into specific position
+    by specifying the same position value for start and stop.
+    The repl string can be appended to each string by specifying -1
+    for both start and stop.
+
+    For details, see :cpp:func:`replace_slice`.
+
+    Parameters
+    ----------
+    input : Column
+        The input strings
+    repl : Scalar, default ""
+        String scalar to replace target with.
+    start : size_type, default 0
+        Start position where repl will be added.
+    stop : size_type, default -1
+        End position (exclusive) to use for replacement.
+    Returns
+    -------
+    pylibcudf.Column
+        New string column
+    """
+    cdef unique_ptr[column] c_result
+
+    if repl is None:
+        repl = Scalar.from_libcudf(
+            cpp_make_string_scalar("".encode())
+        )
+
+    cdef const string_scalar* scalar_str = <string_scalar*>(repl.c_obj.get())
+
+    with nogil:
+        c_result = move(cpp_replace_slice(
+            input.view(),
+            scalar_str[0],
+            start,
+            stop
+        ))
+
+    return Column.from_libcudf(move(c_result))
diff --git a/python/cudf/cudf/_lib/strings/replace.pyx b/python/cudf/cudf/_lib/strings/replace.pyx
index 2d9330a8a24..374831f1833 100644
--- a/python/cudf/cudf/_lib/strings/replace.pyx
+++ b/python/cudf/cudf/_lib/strings/replace.pyx
@@ -1,23 +1,15 @@
 # Copyright (c) 2020-2024, NVIDIA CORPORATION.
 
 from libc.stdint cimport int32_t
-from libcpp.memory cimport unique_ptr
-from libcpp.utility cimport move
 
 from cudf.core.buffer import acquire_spill_lock
 
 from cudf._lib.column cimport Column
-from cudf._lib.pylibcudf.libcudf.column.column cimport column
-from cudf._lib.pylibcudf.libcudf.column.column_view cimport column_view
-from cudf._lib.pylibcudf.libcudf.scalar.scalar cimport string_scalar
-from cudf._lib.pylibcudf.libcudf.strings.replace cimport (
-    replace as cpp_replace,
-    replace_multiple as cpp_replace_multiple,
-    replace_slice as cpp_replace_slice,
-)
 from cudf._lib.pylibcudf.libcudf.types cimport size_type
 from cudf._lib.scalar cimport DeviceScalar
 
+import cudf._lib.pylibcudf as plc
+
 
 @acquire_spill_lock()
 def slice_replace(Column source_strings,
@@ -32,22 +24,12 @@ def slice_replace(Column source_strings,
 
     cdef DeviceScalar repl = py_repl.device_value
 
-    cdef unique_ptr[column] c_result
-    cdef column_view source_view = source_strings.view()
-
-    cdef const string_scalar* scalar_str = <const string_scalar*>(
-        repl.get_raw_ptr()
-    )
-
-    with nogil:
-        c_result = move(cpp_replace_slice(
-            source_view,
-            scalar_str[0],
-            start,
-            stop
-        ))
-
-    return Column.from_unique_ptr(move(c_result))
+    return Column.from_pylibcudf(plc.strings.replace.replace_slice(
+        source_strings.to_pylibcudf(mode="read"),
+        repl.c_value,
+        start,
+        stop
+    ))
 
 
 @acquire_spill_lock()
@@ -61,22 +43,12 @@ def insert(Column source_strings,
 
     cdef DeviceScalar repl = py_repl.device_value
 
-    cdef unique_ptr[column] c_result
-    cdef column_view source_view = source_strings.view()
-
-    cdef const string_scalar* scalar_str = <const string_scalar*>(
-        repl.get_raw_ptr()
-    )
-
-    with nogil:
-        c_result = move(cpp_replace_slice(
-            source_view,
-            scalar_str[0],
-            start,
-            start
-        ))
-
-    return Column.from_unique_ptr(move(c_result))
+    return Column.from_pylibcudf(plc.strings.replace.replace_slice(
+        source_strings.to_pylibcudf(mode="read"),
+        repl.c_value,
+        start,
+        start,
+    ))
 
 
 @acquire_spill_lock()
@@ -92,25 +64,12 @@ def replace(Column source_strings,
     cdef DeviceScalar target = py_target.device_value
     cdef DeviceScalar repl = py_repl.device_value
 
-    cdef unique_ptr[column] c_result
-    cdef column_view source_view = source_strings.view()
-
-    cdef const string_scalar* scalar_target = <const string_scalar*>(
-        target.get_raw_ptr()
-    )
-    cdef const string_scalar* scalar_repl = <const string_scalar*>(
-        repl.get_raw_ptr()
-    )
-
-    with nogil:
-        c_result = move(cpp_replace(
-            source_view,
-            scalar_target[0],
-            scalar_repl[0],
-            maxrepl
-        ))
-
-    return Column.from_unique_ptr(move(c_result))
+    return Column.from_pylibcudf(plc.strings.replace.replace(
+        source_strings.to_pylibcudf(mode="read"),
+        target.c_value,
+        repl.c_value,
+        maxrepl
+    ))
 
 
 @acquire_spill_lock()
@@ -121,16 +80,8 @@ def replace_multi(Column source_strings,
     Returns a Column after replacing occurrences of
     patterns `target_strings` with `repl_strings` in `source_strings`.
     """
-    cdef unique_ptr[column] c_result
-    cdef column_view source_view = source_strings.view()
-    cdef column_view target_view = target_strings.view()
-    cdef column_view repl_view = repl_strings.view()
-
-    with nogil:
-        c_result = move(cpp_replace_multiple(
-            source_view,
-            target_view,
-            repl_view
-        ))
-
-    return Column.from_unique_ptr(move(c_result))
+    return Column.from_pylibcudf(plc.strings.replace.replace_multiple(
+        source_strings.to_pylibcudf(mode="read"),
+        target_strings.to_pylibcudf(mode="read"),
+        repl_strings.to_pylibcudf(mode="read"),
+    ))
diff --git a/python/cudf/cudf/pylibcudf_tests/test_string_replace.py b/python/cudf/cudf/pylibcudf_tests/test_string_replace.py
new file mode 100644
index 00000000000..f20edf6a506
--- /dev/null
+++ b/python/cudf/cudf/pylibcudf_tests/test_string_replace.py
@@ -0,0 +1,126 @@
+# Copyright (c) 2024, NVIDIA CORPORATION.
+
+import pyarrow as pa
+import pytest
+from utils import assert_column_eq
+
+import cudf._lib.pylibcudf as plc
+
+
+@pytest.fixture(scope="module")
+def data_col():
+    pa_data_col = pa.array(
+        ["a", "c", "A", "aa", None, "aaaaaaaaa", "AAAA", "ÁÁÁÁ"],
+        type=pa.string(),
+    )
+    return pa_data_col, plc.interop.from_arrow(pa_data_col)
+
+
+@pytest.fixture(scope="module", params=["a", "c", "A", "Á", "aa", "ÁÁÁ"])
+def scalar_repl_target(request):
+    pa_target = pa.scalar(request.param, type=pa.string())
+    return request.param, plc.interop.from_arrow(pa_target)
+
+
+@pytest.fixture(scope="module", params=["b", "B", "", "B́"])
+def scalar_repl(request):
+    pa_repl = pa.scalar(request.param, type=pa.string())
+    return request.param, plc.interop.from_arrow(pa_repl)
+
+
+@pytest.fixture(
+    scope="module",
+    params=[
+        ["a", "c", "A", "ÁÁÁÁ"],
+    ],
+)
+def col_repl_target(request):
+    pa_target = pa.array(request.param, type=pa.string())
+    return (pa_target, plc.interop.from_arrow(pa_target))
+
+
+@pytest.fixture(
+    scope="module",
+    params=[
+        [
+            "",
+            "z",
+            "XX",
+            "blahblah",
+        ]
+    ],
+)
+def col_repl(request):
+    pa_repl = pa.array(request.param, type=pa.string())
+    return (pa_repl, plc.interop.from_arrow(pa_repl))
+
+
+@pytest.mark.parametrize("maxrepl", [-1, 1, 2, 10])
+def test_replace(data_col, scalar_repl_target, scalar_repl, maxrepl):
+    pa_data_col, plc_data_col = data_col
+    pa_target, plc_target = scalar_repl_target
+    pa_repl, plc_repl = scalar_repl
+    got = plc.strings.replace.replace(
+        plc_data_col, plc_target, plc_repl, maxrepl
+    )
+
+    expected = pa.compute.replace_substring(
+        pa_data_col,
+        pattern=pa_target,
+        replacement=pa_repl,
+        max_replacements=maxrepl,
+    )
+
+    assert_column_eq(expected, got)
+
+
+@pytest.mark.parametrize("startstop", [(0, -1), (0, 0), (1, 3)])
+def test_replace_slice(data_col, scalar_repl, startstop):
+    pa_data_col, plc_data_col = data_col
+    pa_repl, plc_repl = scalar_repl
+    start, stop = startstop
+    got = plc.strings.replace.replace_slice(
+        plc_data_col, plc_repl, start, stop
+    )
+
+    if stop == -1:
+        # pyarrow doesn't support -1 as stop, so just set to really big number
+
+        # TODO: once libcudf's count_characters() is migrated, we can call
+        # count_characters on the input, take the max and set stop to that
+        stop = 1000
+
+    expected = pa.compute.utf8_replace_slice(pa_data_col, start, stop, pa_repl)
+
+    assert_column_eq(expected, got)
+
+
+def test_replace_col(data_col, col_repl_target, col_repl):
+    pa_data_col, plc_data_col = data_col
+    pa_target, plc_target = col_repl_target
+    pa_repl, plc_repl = col_repl
+    got = plc.strings.replace.replace_multiple(
+        plc_data_col, plc_target, plc_repl
+    )
+
+    # There's nothing in pyarrow that does string replace with columns
+    # for targets/repls, so let's implement our own in python
+
+    def replace_list(elem, targets, repls):
+        for target, repl in zip(targets, repls):
+            res = elem.replace(target, repl)
+            if res != elem:
+                return res
+
+    targets = pa_target.to_pylist()
+    repls = pa_repl.to_pylist()
+
+    expected = pa.array(
+        [
+            replace_list(elem, targets, repls) if elem is not None else None
+            for elem in pa_data_col.to_pylist()
+        ],
+        type=pa.string(),
+    )
+
+    assert_column_eq(expected, got)