diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
index 7429365..8bd7e07 100644
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -35,6 +35,6 @@ jobs:
       run: |
         python -m build
     - name: Install package
-      run: python -m pip install --find-links=dist --no-index --ignore-installed docstring_to_markdown
+      run: python -m pip install --find-links=dist --ignore-installed docstring_to_markdown
     - name: Pip check
       run: python -m pip check
diff --git a/README.md b/README.md
index 94231e4..97a3703 100644
--- a/README.md
+++ b/README.md
@@ -6,7 +6,7 @@
 
 On the fly conversion of Python docstrings to markdown
 
-- Python 3.6+ (tested on 3.8 up to 3.13)
+- Python 3.7+ (tested on 3.8 up to 3.13)
 - can recognise reStructuredText and convert multiple of its features to Markdown
 - since v0.13 includes initial support for Google-formatted docstrings
 
@@ -35,6 +35,11 @@ Traceback (most recent call last):
 docstring_to_markdown.UnknownFormatError
 ```
 
+### Extensibility
+
+`docstring_to_markdown` entry point group allows to add custom converters which follow the `Converter` protocol.
+The built-in converters can be customized by providing entry point with matching name.
+
 ### Development
 
 ```bash
diff --git a/docstring_to_markdown/__init__.py b/docstring_to_markdown/__init__.py
index 62fa804..a9367f0 100644
--- a/docstring_to_markdown/__init__.py
+++ b/docstring_to_markdown/__init__.py
@@ -1,27 +1,59 @@
-from .cpython import cpython_to_markdown
-from .google import google_to_markdown, looks_like_google
-from .plain import looks_like_plain_text, plain_text_to_markdown
-from .rst import looks_like_rst, rst_to_markdown
+from importlib_metadata import entry_points
+from typing import List, TYPE_CHECKING
 
-__version__ = "0.15"
+from .types import Converter
+
+if TYPE_CHECKING:
+    from importlib_metadata import EntryPoint
+
+__version__ = "0.16"
 
 
 class UnknownFormatError(Exception):
     pass
 
 
-def convert(docstring: str) -> str:
-    if looks_like_rst(docstring):
-        return rst_to_markdown(docstring)
+def _entry_points_sort_key(entry_point: 'EntryPoint'):
+    if entry_point.dist is None:
+        return 1
+    if entry_point.dist.name == "docstring-to-markdown":
+        return 0
+    return 1
+
+
+def _load_converters() -> List[Converter]:
+    converter_entry_points = entry_points(
+        group="docstring_to_markdown"
+    )
+    # sort so that the default ones can be overridden
+    sorted_entry_points = sorted(
+        converter_entry_points,
+        key=_entry_points_sort_key
+    )
+    # de-duplicate
+    unique_entry_points = {}
+    for entry_point in sorted_entry_points:
+        unique_entry_points[entry_point.name] = entry_point
 
-    if looks_like_google(docstring):
-        return google_to_markdown(docstring)
+    converters = []
+    for entry_point in unique_entry_points.values():
+        converter_class = entry_point.load()
+        converters.append(converter_class())
 
-    if looks_like_plain_text(docstring):
-        return plain_text_to_markdown(docstring)
+    converters.sort(key=lambda converter: -converter.priority)
 
-    cpython = cpython_to_markdown(docstring)
-    if cpython:
-        return cpython
+    return converters
+
+
+_CONVERTERS = None
+
+
+def convert(docstring: str) -> str:
+    global _CONVERTERS
+    if _CONVERTERS is None:
+        _CONVERTERS = _load_converters()
+    for converter in _CONVERTERS:
+        if converter.can_convert(docstring):
+            return converter.convert(docstring)
 
     raise UnknownFormatError()
diff --git a/docstring_to_markdown/cpython.py b/docstring_to_markdown/cpython.py
index e2cae78..974ea60 100644
--- a/docstring_to_markdown/cpython.py
+++ b/docstring_to_markdown/cpython.py
@@ -1,8 +1,10 @@
 from typing import Union, List
 from re import fullmatch
 
+from .types import Converter
 from ._utils import escape_markdown
 
+
 def _is_cpython_signature_line(line: str) -> bool:
     """CPython uses signature lines in the following format:
 
@@ -30,8 +32,29 @@ def cpython_to_markdown(text: str) -> Union[str, None]:
         escape_markdown('\n'.join(other_lines))
     ])
 
+
 def looks_like_cpython(text: str) -> bool:
     return cpython_to_markdown(text) is not None
 
 
-__all__ = ['looks_like_cpython', 'cpython_to_markdown']
+class CPythonConverter(Converter):
+
+    priority = 10
+
+    def __init__(self) -> None:
+        self._last_docstring: Union[str, None] = None
+        self._converted: Union[str, None] = None
+
+    def can_convert(self, docstring):
+        self._last_docstring = docstring
+        self._converted = cpython_to_markdown(docstring)
+        return self._converted is not None
+
+    def convert(self, docstring):
+        if docstring != self._last_docstring:
+            self._last_docstring = docstring
+            self._converted = cpython_to_markdown(docstring)
+        return self._converted
+
+
+__all__ = ['looks_like_cpython', 'cpython_to_markdown', 'CPythonConverter']
diff --git a/docstring_to_markdown/google.py b/docstring_to_markdown/google.py
index e3ca5a8..156f3da 100644
--- a/docstring_to_markdown/google.py
+++ b/docstring_to_markdown/google.py
@@ -2,6 +2,9 @@
 from textwrap import dedent
 from typing import List
 
+from .types import Converter
+
+
 # All possible sections in Google style docstrings
 SECTION_HEADERS: List[str] = [
     "Args",
@@ -169,3 +172,14 @@ def google_to_markdown(text: str, extract_signature: bool = True) -> str:
     docstring = GoogleDocstring(text)
 
     return docstring.as_markdown()
+
+
+class GoogleConverter(Converter):
+
+    priority = 75
+
+    def can_convert(self, docstring):
+        return looks_like_google(docstring)
+
+    def convert(self, docstring):
+        return google_to_markdown(docstring)
diff --git a/docstring_to_markdown/plain.py b/docstring_to_markdown/plain.py
index d3bf7fb..3c42253 100644
--- a/docstring_to_markdown/plain.py
+++ b/docstring_to_markdown/plain.py
@@ -1,4 +1,5 @@
 from re import fullmatch
+from .types import Converter
 from ._utils import escape_markdown
 
 
@@ -24,4 +25,16 @@ def looks_like_plain_text(value: str) -> bool:
 def plain_text_to_markdown(text: str) -> str:
     return escape_markdown(text)
 
-__all__ = ['looks_like_plain_text', 'plain_text_to_markdown']
+
+class PlainTextConverter(Converter):
+
+    priority = 50
+
+    def can_convert(self, docstring):
+        return looks_like_plain_text(docstring)
+
+    def convert(self, docstring):
+        return plain_text_to_markdown(docstring)
+
+
+__all__ = ['looks_like_plain_text', 'plain_text_to_markdown', 'PlainTextConverter']
diff --git a/docstring_to_markdown/rst.py b/docstring_to_markdown/rst.py
index ba50d91..4149773 100644
--- a/docstring_to_markdown/rst.py
+++ b/docstring_to_markdown/rst.py
@@ -4,6 +4,8 @@
 from typing import Callable, Match, Union, List, Dict
 import re
 
+from .types import Converter
+
 
 class Directive:
     def __init__(
@@ -532,7 +534,7 @@ def _consume_row(self, line: str):
                 self._rows.append(self._split(line))
             self._expecting_row_content = not self._expecting_row_content
         else:
-            self._state += 1
+            self._state += 1    # pragma: no cover
 
 
 class BlockParser(IParser):
@@ -651,11 +653,13 @@ def initiate_parsing(self, line: str, current_language: str):
         if line.strip() == '.. autosummary::':
             language = ''
             line = ''
+            suffix = ''
         else:
             line = re.sub(r'::$', '', line)
+            suffix = '\n\n'
 
         self._start_block(language)
-        return IBlockBeginning(remainder=line.rstrip() + '\n\n')
+        return IBlockBeginning(remainder=line.rstrip() + suffix)
 
 
 class MathBlockParser(IndentedBlockParser):
@@ -825,3 +829,14 @@ def flush_buffer():
     if active_parser:
         markdown += active_parser.finish_consumption(True)
     return markdown
+
+
+class ReStructuredTextConverter(Converter):
+
+    priority = 100
+
+    def can_convert(self, docstring):
+        return looks_like_rst(docstring)
+
+    def convert(self, docstring):
+        return rst_to_markdown(docstring)
diff --git a/docstring_to_markdown/types.py b/docstring_to_markdown/types.py
new file mode 100644
index 0000000..ef7fb18
--- /dev/null
+++ b/docstring_to_markdown/types.py
@@ -0,0 +1,14 @@
+from typing_extensions import Protocol
+
+
+class Converter(Protocol):
+
+    def convert(self, docstring: str) -> str:
+        """Convert given docstring to markdown."""
+
+    def can_convert(self, docstring: str) -> bool:
+        """Check if conversion to markdown can be performed."""
+
+    # The higher the priority, the sooner the conversion
+    # with this converter will be attempted.
+    priority: int
diff --git a/setup.cfg b/setup.cfg
index 0bb013f..817ea9e 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -28,21 +28,36 @@ warn_unused_configs = True
 
 [options]
 packages = find:
-python_requires = >=3.6
+python_requires = >=3.7
+install_requires =
+    importlib-metadata>=3.6
+    typing_extensions>=4.6
 
 [options.package_data]
 docstring-to-markdown = py.typed
 
+[options.entry_points]
+docstring_to_markdown =
+  rst = docstring_to_markdown.rst:ReStructuredTextConverter
+  google = docstring_to_markdown.google:GoogleConverter
+  plain = docstring_to_markdown.plain:PlainTextConverter
+  cpython = docstring_to_markdown.cpython:CPythonConverter
+
 [tool:pytest]
 addopts =
     --pyargs tests
     --cov docstring_to_markdown
-    --cov-fail-under=99
+    --cov-fail-under=100
     --cov-report term-missing:skip-covered
     -p no:warnings
     --flake8
     -vv
 
+[coverage:report]
+exclude_lines =
+    pragma: no cover
+    if TYPE_CHECKING:
+
 [flake8]
 max-line-length = 120
 max-complexity = 15
diff --git a/tests/test_convert.py b/tests/test_convert.py
index 7642aeb..5b030da 100644
--- a/tests/test_convert.py
+++ b/tests/test_convert.py
@@ -1,4 +1,10 @@
+from contextlib import contextmanager
 from docstring_to_markdown import convert, UnknownFormatError
+from docstring_to_markdown.types import Converter
+from docstring_to_markdown.cpython import CPythonConverter
+from importlib_metadata import EntryPoint, entry_points, distribution
+from unittest.mock import patch
+import docstring_to_markdown
 import pytest
 
 CPYTHON = """\
@@ -55,3 +61,72 @@ def test_convert_rst():
 def test_unknown_format():
     with pytest.raises(UnknownFormatError):
         convert('ARGS [arg1, arg2] RETURNS: str OR None')
+
+
+class HighPriorityConverter(Converter):
+    priority = 120
+
+    def convert(self, docstring):
+        return "HighPriority"
+
+    def can_convert(self, docstring):
+        return True
+
+
+class MockEntryPoint(EntryPoint):
+    def load(self):
+        return self.value
+
+    dist = None
+
+
+class DistMockEntryPoint(MockEntryPoint):
+    # Pretend it is contributed by `pytest`.
+    # It could be anything else, but `pytest`
+    # is guaranteed to be installed during tests.
+    dist = distribution('pytest')
+
+
+class CustomCPythonConverter(CPythonConverter):
+    priority = 10
+
+    def convert(self, docstring):
+        return 'CustomCPython'
+
+    def can_convert(self, docstring):
+        return True
+
+
+@contextmanager
+def custom_entry_points(entry_points):
+    old = docstring_to_markdown._CONVERTERS
+    docstring_to_markdown._CONVERTERS = None
+    with patch.object(docstring_to_markdown, 'entry_points', return_value=entry_points):
+        yield
+    docstring_to_markdown._CONVERTERS = old
+
+
+def test_adding_entry_point():
+    original_entry_points = entry_points(group="docstring_to_markdown")
+    mock_entry_point = MockEntryPoint(
+        name='high-priority-converter',
+        group='docstring_to_markdown',
+        value=HighPriorityConverter,
+    )
+    with custom_entry_points([*original_entry_points, mock_entry_point]):
+        assert convert('test') == 'HighPriority'
+
+
+def test_replacing_entry_point():
+    assert convert(CPYTHON) == CPYTHON_MD
+    original_entry_points = entry_points(group="docstring_to_markdown")
+    mock_entry_point = DistMockEntryPoint(
+        name='cpython',
+        group='docstring_to_markdown',
+        value=CustomCPythonConverter
+    )
+    with custom_entry_points([*original_entry_points, mock_entry_point]):
+        assert convert('test') == 'test'
+        assert convert(GOOGLE) == GOOGLE_MD
+        assert convert(RST) == RST_MD
+        assert convert(CPYTHON) == 'CustomCPython'
diff --git a/tests/test_cpython.py b/tests/test_cpython.py
index 2b7245d..b292c66 100644
--- a/tests/test_cpython.py
+++ b/tests/test_cpython.py
@@ -1,5 +1,5 @@
 import pytest
-from docstring_to_markdown.cpython import looks_like_cpython, cpython_to_markdown
+from docstring_to_markdown.cpython import looks_like_cpython, cpython_to_markdown, CPythonConverter
 
 BOOL = """\
 bool(x) -> bool
@@ -101,3 +101,10 @@ def test_conversion_bool():
 
 def test_conversion_str():
     assert cpython_to_markdown(STR) == STR_MD
+
+
+def test_convert():
+    converter = CPythonConverter()
+    assert converter.can_convert(BOOL)
+    assert not converter.can_convert('this is plain text')
+    assert converter.convert(BOOL) == BOOL_MD
diff --git a/tests/test_google.py b/tests/test_google.py
index 78486bc..ef0d785 100644
--- a/tests/test_google.py
+++ b/tests/test_google.py
@@ -1,6 +1,6 @@
 import pytest
 
-from docstring_to_markdown.google import google_to_markdown, looks_like_google
+from docstring_to_markdown.google import google_to_markdown, looks_like_google, GoogleConverter
 
 BASIC_EXAMPLE = """Do **something**.
 
@@ -133,3 +133,10 @@ def test_looks_like_google_ignores_plain_text():
 )
 def test_google_to_markdown(google, markdown):
     assert google_to_markdown(google) == markdown
+
+
+def test_converter():
+    converter = GoogleConverter()
+    assert converter.can_convert(BASIC_EXAMPLE)
+    assert not converter.can_convert("This is plain text")
+    assert converter.convert(BASIC_EXAMPLE) == BASIC_EXAMPLE_MD
diff --git a/tests/test_plain.py b/tests/test_plain.py
index 61ef4b4..0e30eef 100644
--- a/tests/test_plain.py
+++ b/tests/test_plain.py
@@ -1,5 +1,5 @@
 import pytest
-from docstring_to_markdown.plain import looks_like_plain_text, plain_text_to_markdown
+from docstring_to_markdown.plain import looks_like_plain_text, plain_text_to_markdown, PlainTextConverter
 
 
 @pytest.mark.parametrize("text", [
@@ -40,3 +40,10 @@ def test_rejects_code(text):
 
 def test_conversion():
     assert plain_text_to_markdown("test") == "test"
+
+
+def test_convert():
+    converter = PlainTextConverter()
+    assert not converter.can_convert('def test():')
+    assert converter.can_convert('this is plain text')
+    assert converter.convert('test') == 'test'
diff --git a/tests/test_rst.py b/tests/test_rst.py
index 3830563..c645def 100644
--- a/tests/test_rst.py
+++ b/tests/test_rst.py
@@ -1,6 +1,6 @@
 import pytest
 
-from docstring_to_markdown.rst import looks_like_rst, rst_to_markdown
+from docstring_to_markdown.rst import looks_like_rst, rst_to_markdown, ReStructuredTextConverter
 
 
 SEE_ALSO = """
@@ -119,6 +119,26 @@
 A function definition is an executable statement.
 """
 
+RST_AUTOSUMMARY_BLOCK = """
+Summary
+
+.. autosummary::
+
+   environment.BuildEnvironment
+   util.relative_uri
+"""
+
+
+RST_AUTOSUMMARY_BLOCK_MARKDOWN = """
+Summary
+
+```
+environment.BuildEnvironment
+util.relative_uri
+```
+"""
+
+
 RST_COLON_CODE_BLOCK = """
 For example, the following code ::
 
@@ -598,6 +618,26 @@ def func(): pass
 """
 
 
+BROKEN_GRID_TABLE = """
++------------+-----------+------------+-----------------+---+---------+
+|param_kernel|param_gamma|param_degree|split0_test_score|...|rank_t...|
++============+===========+============+=================+===+=========+
+|  'poly'    |     --    |      2     |       0.80      |...|    2    |
++------------+-----------+------------+-----------------+---+---------+
+|  'poly'    |     --    |      3     |       0.70      |...|    4    |
+someone forgot to close the row above.
+"""
+
+
+BROKEN_GRID_TABLE_MARKDOWN = """
+| param_kernel | param_gamma | param_degree | split0_test_score | ... | rank_t... |
+| ------------ | ----------- | ------------ | ----------------- | --- | --------- |
+| 'poly'       | --          | 2            | 0.80              | ... | 2         |
+| 'poly'       | --          | 3            | 0.70              | ... | 4         |
+someone forgot to close the row above.
+"""
+
+
 NESTED_PARAMETERS = """
 Parameters
 ----------
@@ -733,6 +773,10 @@ def foo():
         'rst': NUMPY_EXAMPLE,
         'md': NUMPY_EXAMPLE_MARKDOWN
     },
+    'converts autosummary block': {
+        'rst': RST_AUTOSUMMARY_BLOCK,
+        'md': RST_AUTOSUMMARY_BLOCK_MARKDOWN
+    },
     'converts version changed': {
         'rst': '.. versionchanged:: 0.23.0',
         'md': '*Changed in 0.23.0*'
@@ -835,6 +879,10 @@ def foo():
         'rst': GRID_TABLE_IN_SKLEARN,
         'md': GRID_TABLE_IN_SKLEARN_MARKDOWN
     },
+    'converts broken grid table': {
+        'rst': BROKEN_GRID_TABLE,
+        'md': BROKEN_GRID_TABLE_MARKDOWN
+    },
     'converts nested parameter lists': {
         'rst': NESTED_PARAMETERS,
         'md': NESTED_PARAMETERS_MARKDOWN
@@ -888,3 +936,10 @@ def test_rst_to_markdown(rst, markdown):
     converted = rst_to_markdown(rst)
     print(converted)
     assert converted == markdown
+
+
+def test_converter():
+    converter = ReStructuredTextConverter()
+    assert converter.can_convert('.. versionadded:: 0.1')
+    assert not converter.can_convert('this is plain text')
+    assert converter.convert(PEP_287_CODE_BLOCK) == PEP_287_CODE_BLOCK_MARKDOWN