diff --git a/setup.cfg b/setup.cfg
index 0aa2212..9427f05 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -74,7 +74,7 @@ tests =
     pytest
     requests_file
 docs =
-    sphinx
+    sphinx<8.0
     sphinx-rtd-theme
     sphinx-click
     sphinx-autodoc-typehints
diff --git a/src/pystow/__main__.py b/src/pystow/__main__.py
index 0405cdf..350e058 100644
--- a/src/pystow/__main__.py
+++ b/src/pystow/__main__.py
@@ -1,4 +1,5 @@
 # -*- coding: utf-8 -*-
+# type: ignore
 
 """Command line interface for PyStow."""
 
diff --git a/src/pystow/api.py b/src/pystow/api.py
index fb16837..44fee3a 100644
--- a/src/pystow/api.py
+++ b/src/pystow/api.py
@@ -2,11 +2,12 @@
 
 """API functions for PyStow."""
 
+import sqlite3
 from contextlib import contextmanager
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, Mapping, Optional, Sequence, Union
+from typing import TYPE_CHECKING, Any, Generator, Mapping, Optional, Sequence, Union
 
-from .constants import JSON, Opener, Provider
+from .constants import JSON, BytesOpener, Opener, Provider
 from .impl import Module
 
 if TYPE_CHECKING:
@@ -115,7 +116,7 @@ def open(
     name: str,
     mode: str = "r",
     open_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> Opener:
     """Open a file that exists already.
 
     :param key:
@@ -143,7 +144,7 @@ def open_gz(
     name: str,
     mode: str = "rt",
     open_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> Opener:
     """Open a gzipped file that exists already.
 
     :param key:
@@ -205,7 +206,7 @@ def ensure_custom(
     name: str,
     force: bool = False,
     provider: Provider,
-    **kwargs,
+    **kwargs: Any,
 ) -> Path:
     """Ensure a file is present, and run a custom create function otherwise.
 
@@ -382,7 +383,7 @@ def ensure_open_zip(
     download_kwargs: Optional[Mapping[str, Any]] = None,
     mode: str = "r",
     open_kwargs: Optional[Mapping[str, Any]] = None,
-) -> Opener:
+) -> BytesOpener:
     """Ensure a file is downloaded then open it with :mod:`zipfile`.
 
     :param key:
@@ -480,7 +481,7 @@ def ensure_open_tarfile(
     download_kwargs: Optional[Mapping[str, Any]] = None,
     mode: str = "r",
     open_kwargs: Optional[Mapping[str, Any]] = None,
-) -> Opener:
+) -> BytesOpener:
     """Ensure a tar file is downloaded and open a file inside it.
 
     :param key:
@@ -708,7 +709,7 @@ def dump_df(
     name: str,
     obj: "pd.DataFrame",
     sep: str = "\t",
-    index=False,
+    index: bool = False,
     to_csv_kwargs: Optional[Mapping[str, Any]] = None,
 ) -> None:
     """Dump a dataframe to a TSV file with :mod:`pandas`.
@@ -1197,7 +1198,7 @@ def ensure_tar_df(
     force: bool = False,
     download_kwargs: Optional[Mapping[str, Any]] = None,
     read_csv_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> "pd.DataFrame":
     """Download a tar file and open an inner file as a dataframe with :mod:`pandas`.
 
     :param key: The module name
@@ -1241,7 +1242,7 @@ def ensure_tar_xml(
     force: bool = False,
     download_kwargs: Optional[Mapping[str, Any]] = None,
     parse_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> "lxml.etree.ElementTree":
     """Download a tar file and open an inner file as an XML with :mod:`lxml`.
 
     :param key: The module name
@@ -1453,7 +1454,7 @@ def dump_rdf(
     obj: "rdflib.Graph",
     format: str = "turtle",
     serialize_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> None:
     """Dump an RDF graph to a file with :mod:`rdflib`.
 
     :param key:
@@ -1480,7 +1481,7 @@ def ensure_from_s3(
     s3_key: Union[str, Sequence[str]],
     name: Optional[str] = None,
     force: bool = False,
-    **kwargs,
+    **kwargs: Any,
 ) -> Path:
     """Ensure a file is downloaded.
 
@@ -1512,7 +1513,7 @@ def ensure_from_s3(
     """
     _module = Module.from_key(key, ensure_exists=True)
     return _module.ensure_from_s3(
-        *subkeys, s3_bucket=s3_bucket, s3_key=s3_key, name=name, force=force
+        *subkeys, s3_bucket=s3_bucket, s3_key=s3_key, name=name, force=force, **kwargs
     )
 
 
@@ -1578,7 +1579,7 @@ def ensure_open_sqlite(
     name: Optional[str] = None,
     force: bool = False,
     download_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> Generator[sqlite3.Connection, None, None]:
     """Ensure and connect to a SQLite database.
 
     :param key:
@@ -1622,7 +1623,7 @@ def ensure_open_sqlite_gz(
     name: Optional[str] = None,
     force: bool = False,
     download_kwargs: Optional[Mapping[str, Any]] = None,
-):
+) -> Generator[sqlite3.Connection, None, None]:
     """Ensure and connect to a gzipped SQLite database.
 
     :param key:
diff --git a/src/pystow/cache.py b/src/pystow/cache.py
index 04bff89..83651dd 100644
--- a/src/pystow/cache.py
+++ b/src/pystow/cache.py
@@ -5,7 +5,6 @@
 import functools
 import json
 import logging
-import os
 from abc import ABC, abstractmethod
 from pathlib import Path
 from typing import (
@@ -25,7 +24,7 @@
 try:
     import pickle5 as pickle
 except ImportError:
-    import pickle  # type:ignore
+    import pickle
 
 if TYPE_CHECKING:
     import pandas as pd
@@ -57,7 +56,7 @@ class Cached(Generic[X], ABC):
 
     def __init__(
         self,
-        path: Union[str, Path, os.PathLike],
+        path: Union[str, Path],
         force: bool = False,
     ) -> None:
         """Instantiate the decorator.
@@ -165,7 +164,7 @@ class CachedDataFrame(Cached["pd.DataFrame"]):
 
     def __init__(
         self,
-        path: Union[str, Path, os.PathLike],
+        path: Union[str, Path],
         force: bool = False,
         sep: Optional[str] = None,
         dtype: Optional[Any] = None,
diff --git a/src/pystow/cli.py b/src/pystow/cli.py
index 55fb618..2de3bfb 100644
--- a/src/pystow/cli.py
+++ b/src/pystow/cli.py
@@ -1,5 +1,6 @@
 # -*- coding: utf-8 -*-
 # flake8: noqa
+# type: ignore
 
 """Command line interface for PyStow."""
 
@@ -10,7 +11,7 @@
 
 
 @click.group()
-def main():
+def main() -> None:
     """Run the PyStow CLI."""
 
 
diff --git a/src/pystow/config_api.py b/src/pystow/config_api.py
index 91fa599..34d92b4 100644
--- a/src/pystow/config_api.py
+++ b/src/pystow/config_api.py
@@ -7,7 +7,7 @@
 from functools import lru_cache
 from pathlib import Path
 from textwrap import dedent
-from typing import Optional, Type, TypeVar
+from typing import Any, Callable, Optional, Type, TypeVar, Union
 
 from .utils import getenv_path
 
@@ -131,7 +131,7 @@ def get_config(
     default: Optional[X] = None,
     dtype: Optional[Type[X]] = None,
     raise_on_missing: bool = False,
-):
+) -> Any:
     """Get a configuration value.
 
     :param module: Name of the module (e.g., ``pybel``) to get configuration for
@@ -159,13 +159,13 @@ def get_config(
     return _cast(rv, dtype)
 
 
-def _cast(rv, dtype):
+def _cast(rv: Any, dtype: Union[None, Callable[..., Any]]) -> Any:
     if not isinstance(rv, str):  # if it's not a string, it doesn't need munging
         return rv
     if dtype in (None, str):  # no munging necessary
         return rv
     if dtype in (int, float):
-        return dtype(rv)  # type: ignore
+        return dtype(rv)
     if dtype is bool:
         if rv.lower() in ("t", "true", "yes", "1", 1, True):
             return True
diff --git a/src/pystow/constants.py b/src/pystow/constants.py
index 719fb58..408a77c 100644
--- a/src/pystow/constants.py
+++ b/src/pystow/constants.py
@@ -2,8 +2,9 @@
 
 """PyStow constants."""
 
+from io import StringIO
 from textwrap import dedent
-from typing import IO, Any, Callable, Iterator
+from typing import IO, Any, Callable, Generator
 
 __all__ = [
     "PYSTOW_NAME_ENVVAR",
@@ -68,6 +69,7 @@
 """
 )
 
-Opener = Iterator[IO]
+Opener = Generator[StringIO, None, None]
+BytesOpener = Generator[IO[bytes], None, None]
 JSON = Any
 Provider = Callable[..., None]
diff --git a/src/pystow/impl.py b/src/pystow/impl.py
index 32cc5f0..e6be937 100644
--- a/src/pystow/impl.py
+++ b/src/pystow/impl.py
@@ -7,14 +7,24 @@
 import json
 import logging
 import lzma
+import sqlite3
 import tarfile
 import zipfile
 from contextlib import closing, contextmanager
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, Mapping, Optional, Sequence, Union
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    Generator,
+    Mapping,
+    Optional,
+    Sequence,
+    Union,
+)
 
 from . import utils
-from .constants import JSON, Opener, Provider
+from .constants import JSON, BytesOpener, Opener, Provider
 from .utils import (
     base_from_gzip_name,
     download_from_google,
@@ -35,11 +45,12 @@
 try:
     import pickle5 as pickle
 except ImportError:
-    import pickle  # type:ignore
+    import pickle
 
 if TYPE_CHECKING:
     import botocore.client
     import lxml.etree
+    import numpy
     import pandas as pd
     import rdflib
 
@@ -182,7 +193,7 @@ def ensure_custom(
         name: str,
         force: bool = False,
         provider: Provider,
-        **kwargs,
+        **kwargs: Any,
     ) -> Path:
         """Ensure a file is present, and run a custom create function otherwise.
 
@@ -448,7 +459,7 @@ def ensure_open_tarfile(
         download_kwargs: Optional[Mapping[str, Any]] = None,
         mode: str = "r",
         open_kwargs: Optional[Mapping[str, Any]] = None,
-    ) -> Opener:
+    ) -> BytesOpener:
         """Ensure a tar file is downloaded and open a file inside it.
 
         :param subkeys:
@@ -490,7 +501,7 @@ def ensure_open_zip(
         download_kwargs: Optional[Mapping[str, Any]] = None,
         mode: str = "r",
         open_kwargs: Optional[Mapping[str, Any]] = None,
-    ) -> Opener:
+    ) -> BytesOpener:
         """Ensure a file is downloaded then open it with :mod:`zipfile`.
 
         :param subkeys:
@@ -659,7 +670,7 @@ def dump_df(
         name: str,
         obj: "pd.DataFrame",
         sep: str = "\t",
-        index=False,
+        index: bool = False,
         to_csv_kwargs: Optional[Mapping[str, Any]] = None,
     ) -> None:
         """Dump a dataframe to a TSV file with :mod:`pandas`.
@@ -729,7 +740,7 @@ def ensure_json_bz2(
         download_kwargs: Optional[Mapping[str, Any]] = None,
         open_kwargs: Optional[Mapping[str, Any]] = None,
         json_load_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> JSON:
         """Download BZ2-compressed JSON and open with :mod:`json`.
 
         :param subkeys:
@@ -1102,7 +1113,7 @@ def dump_xml(
         obj: "lxml.etree.ElementTree",
         open_kwargs: Optional[Mapping[str, Any]] = None,
         write_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> None:
         """Dump an XML element tree to a file with :mod:`lxml`.
 
         :param subkeys:
@@ -1127,7 +1138,7 @@ def ensure_tar_xml(
         force: bool = False,
         download_kwargs: Optional[Mapping[str, Any]] = None,
         parse_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> "lxml.etree.ElementTree":
         """Download a tar file and open an inner file as an XML with :mod:`lxml`.
 
         :param subkeys:
@@ -1200,7 +1211,7 @@ def ensure_zip_np(
         force: bool = False,
         download_kwargs: Optional[Mapping[str, Any]] = None,
         load_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> "numpy.typing.ArrayLike":
         """Download a zip file and open an inner file as an array-like with :mod:`numpy`.
 
         :param subkeys:
@@ -1269,11 +1280,11 @@ def ensure_rdf(
         cache_path = path.with_suffix(path.suffix + ".pickle.gz")
         if cache_path.exists() and not force:
             with gzip.open(cache_path, "rb") as file:
-                return pickle.load(file)  # type: ignore
+                return pickle.load(file)
 
         rv = read_rdf(path=path, **(parse_kwargs or {}))
         with gzip.open(cache_path, "wb") as file:
-            pickle.dump(rv, file, protocol=pickle.HIGHEST_PROTOCOL)  # type: ignore
+            pickle.dump(rv, file, protocol=pickle.HIGHEST_PROTOCOL)
         return rv
 
     def load_rdf(
@@ -1303,7 +1314,7 @@ def dump_rdf(
         obj: "rdflib.Graph",
         format: str = "turtle",
         serialize_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> None:
         """Dump an RDF graph to a file with :mod:`rdflib`.
 
         :param subkeys:
@@ -1409,7 +1420,7 @@ def ensure_open_sqlite(
         name: Optional[str] = None,
         force: bool = False,
         download_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> Generator[sqlite3.Connection, None, None]:
         """Ensure and connect to a SQLite database.
 
         :param subkeys:
@@ -1435,8 +1446,6 @@ def ensure_open_sqlite(
         >>> with module.ensure_open_sqlite(url=url) as conn:
         >>>     df = pd.read_sql(sql, conn)
         """
-        import sqlite3
-
         path = self.ensure(
             *subkeys, url=url, name=name, force=force, download_kwargs=download_kwargs
         )
@@ -1451,7 +1460,7 @@ def ensure_open_sqlite_gz(
         name: Optional[str] = None,
         force: bool = False,
         download_kwargs: Optional[Mapping[str, Any]] = None,
-    ):
+    ) -> Generator[sqlite3.Connection, None, None]:
         """Ensure and connect to a SQLite database that's gzipped.
 
         Unfortunately, it's a paid feature to directly read gzipped sqlite files,
@@ -1480,8 +1489,6 @@ def ensure_open_sqlite_gz(
         >>> with module.ensure_open_sqlite_gz(url=url) as conn:
         >>>     df = pd.read_sql(sql, conn)
         """
-        import sqlite3
-
         path = self.ensure_gunzip(
             *subkeys, url=url, name=name, force=force, download_kwargs=download_kwargs
         )
@@ -1489,7 +1496,7 @@ def ensure_open_sqlite_gz(
             yield conn
 
 
-def _clean_csv_kwargs(read_csv_kwargs):
+def _clean_csv_kwargs(read_csv_kwargs: Union[None, Mapping[str, Any]]) -> Dict[str, Any]:
     read_csv_kwargs = {} if read_csv_kwargs is None else dict(read_csv_kwargs)
     read_csv_kwargs.setdefault("sep", "\t")
     return read_csv_kwargs
diff --git a/src/pystow/utils.py b/src/pystow/utils.py
index 801f912..6e2beea 100644
--- a/src/pystow/utils.py
+++ b/src/pystow/utils.py
@@ -18,7 +18,16 @@
 from io import BytesIO, StringIO
 from pathlib import Path, PurePosixPath
 from subprocess import check_output  # noqa: S404
-from typing import Any, Collection, Iterable, Iterator, Mapping, Optional, Union
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Collection,
+    Iterable,
+    Iterator,
+    Mapping,
+    Optional,
+    Union,
+)
 from urllib.parse import urlparse
 from urllib.request import urlretrieve
 from uuid import uuid4
@@ -34,6 +43,13 @@
     README_TEXT,
 )
 
+if TYPE_CHECKING:
+    import botocore.client
+    import lxml.etree
+    import numpy.typing
+    import pandas
+    import rdflib
+
 __all__ = [
     # Data Structures
     "HexDigestMismatch",
@@ -104,7 +120,7 @@ def __init__(self, offending_hexdigests: Collection[HexDigestMismatch]):
         """
         self.offending_hexdigests = offending_hexdigests
 
-    def __str__(self):  # noqa:D105
+    def __str__(self) -> str:  # noqa:D105
         return "\n".join(
             (
                 "Hexdigest of downloaded file does not match the expected ones!",
@@ -228,7 +244,7 @@ def get_hashes(
     # calculate hash sums of file incrementally
     buffer = memoryview(bytearray(chunk_size))
     with path.open("rb", buffering=0) as file:
-        for this_chunk_size in iter(lambda: file.readinto(buffer), 0):  # type: ignore
+        for this_chunk_size in iter(lambda: file.readinto(buffer), 0):
             for alg in algorithms.values():
                 alg.update(buffer[:this_chunk_size])
 
@@ -266,7 +282,7 @@ def raise_on_digest_mismatch(
         raise HexDigestError(offending_hexdigests)
 
 
-class TqdmReportHook(tqdm):
+class TqdmReportHook(tqdm):  # type:ignore
     """A custom progress bar that can be used with urllib.
 
     Based on https://gist.github.com/leimao/37ff6e990b3226c2c9670a2cd1e4a6f5
@@ -372,7 +388,7 @@ def download(
                 # Solution for progres bar from https://stackoverflow.com/a/63831344/5775947
                 total_size = int(response.headers.get("Content-Length", 0))
                 # Decompress if needed
-                response.raw.read = partial(response.raw.read, decode_content=True)
+                response.raw.read = partial(response.raw.read, decode_content=True)  # type:ignore
                 with tqdm.wrapattr(response.raw, "read", total=total_size, **_tqdm_kwargs) as fsrc:
                     shutil.copyfileobj(fsrc, file)
         else:
@@ -484,11 +500,12 @@ def n() -> str:
     return str(uuid4())
 
 
-def get_df_io(df, sep: str = "\t", index: bool = False, **kwargs) -> BytesIO:
+def get_df_io(
+    df: "pandas.DataFrame", sep: str = "\t", index: bool = False, **kwargs: Any
+) -> BytesIO:
     """Get the dataframe as bytes.
 
     :param df: A dataframe
-    :type df: pandas.DataFrame
     :param sep: The separator in the dataframe. Overrides Pandas default to use a tab.
     :param index:  Should the index be output? Overrides the Pandas default to be false.
     :param kwargs: Additional kwargs to pass to :func:`pandas.DataFrame.to_csv`.
@@ -501,7 +518,7 @@ def get_df_io(df, sep: str = "\t", index: bool = False, **kwargs) -> BytesIO:
     return bio
 
 
-def get_np_io(arr, **kwargs) -> BytesIO:
+def get_np_io(arr: "numpy.typing.ArrayLike", **kwargs: Any) -> BytesIO:
     """Get the numpy object as bytes.
 
     :param arr: Array-like
@@ -517,9 +534,9 @@ def get_np_io(arr, **kwargs) -> BytesIO:
 
 
 def write_pickle_gz(
-    obj,
+    obj: Any,
     path: Union[str, Path],
-    **kwargs,
+    **kwargs: Any,
 ) -> None:
     """Write an object to a gzipped pickle.
 
@@ -533,12 +550,12 @@ def write_pickle_gz(
 
 
 def write_lzma_csv(
-    df,
+    df: "pandas.DataFrame",
     path: Union[str, Path],
-    sep="\t",
+    sep: str = "\t",
     index: bool = False,
-    **kwargs,
-):
+    **kwargs: Any,
+) -> None:
     """Write a dataframe as an lzma-compressed file.
 
     :param df: A dataframe
@@ -556,12 +573,12 @@ def write_lzma_csv(
 
 
 def write_zipfile_csv(
-    df,
+    df: "pandas.DataFrame",
     path: Union[str, Path],
     inner_path: str,
-    sep="\t",
+    sep: str = "\t",
     index: bool = False,
-    **kwargs,
+    **kwargs: Any,
 ) -> None:
     """Write a dataframe to an inner CSV file to a zip archive.
 
@@ -581,7 +598,9 @@ def write_zipfile_csv(
             file.write(bytes_io.read())
 
 
-def read_zipfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", **kwargs):
+def read_zipfile_csv(
+    path: Union[str, Path], inner_path: str, sep: str = "\t", **kwargs: Any
+) -> "pandas.DataFrame":
     """Read an inner CSV file from a zip archive.
 
     :param path: The path to the zip archive
@@ -589,7 +608,6 @@ def read_zipfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", *
     :param sep: The separator in the dataframe. Overrides Pandas default to use a tab.
     :param kwargs: Additional kwargs to pass to :func:`pandas.read_csv`.
     :return: A dataframe
-    :rtype: pandas.DataFrame
     """
     import pandas as pd
 
@@ -599,15 +617,14 @@ def read_zipfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", *
 
 
 def write_zipfile_xml(
-    element_tree,
+    element_tree: "lxml.etree.ElementTree",
     path: Union[str, Path],
     inner_path: str,
-    **kwargs,
+    **kwargs: Any,
 ) -> None:
     """Write an XML element tree to an inner XML file to a zip archive.
 
     :param element_tree: An XML element tree
-    :type element_tree: lxml.etree.ElementTree
     :param path: The path to the resulting zip archive
     :param inner_path: The path inside the zip archive to write the dataframe
     :param kwargs: Additional kwargs to pass to :func:`tostring`
@@ -620,14 +637,15 @@ def write_zipfile_xml(
             file.write(etree.tostring(element_tree, **kwargs))
 
 
-def read_zipfile_xml(path: Union[str, Path], inner_path: str, **kwargs):
+def read_zipfile_xml(
+    path: Union[str, Path], inner_path: str, **kwargs: Any
+) -> "lxml.etree.ElementTree":
     """Read an inner XML file from a zip archive.
 
     :param path: The path to the zip archive
     :param inner_path: The path inside the zip archive to the xml file
     :param kwargs: Additional kwargs to pass to :func:`lxml.etree.parse`
     :return: An XML element tree
-    :rtype: lxml.etree.ElementTree
     """
     from lxml import etree
 
@@ -637,10 +655,10 @@ def read_zipfile_xml(path: Union[str, Path], inner_path: str, **kwargs):
 
 
 def write_zipfile_np(
-    arr,
+    arr: "numpy.typing.ArrayLike",
     path: Union[str, Path],
     inner_path: str,
-    **kwargs,
+    **kwargs: Any,
 ) -> None:
     """Write a dataframe to an inner CSV file to a zip archive.
 
@@ -657,14 +675,13 @@ def write_zipfile_np(
             file.write(bytes_io.read())
 
 
-def read_zip_np(path: Union[str, Path], inner_path: str, **kwargs):
+def read_zip_np(path: Union[str, Path], inner_path: str, **kwargs: Any) -> "numpy.typing.ArrayLike":
     """Read an inner numpy array-like from a zip archive.
 
     :param path: The path to the zip archive
     :param inner_path: The path inside the zip archive to the dataframe
     :param kwargs: Additional kwargs to pass to :func:`numpy.load`.
     :return: A numpy array or other object
-    :rtype: numpy.typing.ArrayLike
     """
     import numpy as np
 
@@ -673,14 +690,13 @@ def read_zip_np(path: Union[str, Path], inner_path: str, **kwargs):
             return np.load(file, **kwargs)
 
 
-def read_zipfile_rdf(path: Union[str, Path], inner_path: str, **kwargs):
+def read_zipfile_rdf(path: Union[str, Path], inner_path: str, **kwargs: Any) -> "rdflib.Graph":
     """Read an inner RDF file from a zip archive.
 
     :param path: The path to the zip archive
     :param inner_path: The path inside the zip archive to the dataframe
     :param kwargs: Additional kwargs to pass to :func:`pandas.read_csv`.
     :return: A dataframe
-    :rtype: rdflib.Graph
     """
     import rdflib
 
@@ -692,12 +708,12 @@ def read_zipfile_rdf(path: Union[str, Path], inner_path: str, **kwargs):
 
 
 def write_tarfile_csv(
-    df,
+    df: "pandas.DataFrame",
     path: Union[str, Path],
     inner_path: str,
     sep: str = "\t",
     index: bool = False,
-    **kwargs,
+    **kwargs: Any,
 ) -> None:
     """Write a dataframe to an inner CSV file from a tar archive.
 
@@ -718,7 +734,9 @@ def write_tarfile_csv(
         tar_file.addfile(tarinfo, BytesIO(s.encode("utf-8")))
 
 
-def read_tarfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", **kwargs):
+def read_tarfile_csv(
+    path: Union[str, Path], inner_path: str, sep: str = "\t", **kwargs: Any
+) -> "pandas.DataFrame":
     """Read an inner CSV file from a tar archive.
 
     :param path: The path to the tar archive
@@ -726,7 +744,6 @@ def read_tarfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", *
     :param sep: The separator in the dataframe. Overrides Pandas default to use a tab.
     :param kwargs: Additional kwargs to pass to :func:`pandas.read_csv`.
     :return: A dataframe
-    :rtype: pandas.DataFrame
     """
     import pandas as pd
 
@@ -735,7 +752,9 @@ def read_tarfile_csv(path: Union[str, Path], inner_path: str, sep: str = "\t", *
             return pd.read_csv(file, sep=sep, **kwargs)
 
 
-def read_tarfile_xml(path: Union[str, Path], inner_path: str, **kwargs):
+def read_tarfile_xml(
+    path: Union[str, Path], inner_path: str, **kwargs: Any
+) -> "lxml.etree.ElementTree":
     """Read an inner XML file from a tar archive.
 
     :param path: The path to the tar archive
@@ -751,13 +770,12 @@ def read_tarfile_xml(path: Union[str, Path], inner_path: str, **kwargs):
             return etree.parse(file, **kwargs)
 
 
-def read_rdf(path: Union[str, Path], **kwargs):
+def read_rdf(path: Union[str, Path], **kwargs: Any) -> "rdflib.Graph":
     """Read an RDF file with :mod:`rdflib`.
 
     :param path: The path to the RDF file
     :param kwargs: Additional kwargs to pass to :func:`rdflib.Graph.parse`
     :return: A parsed RDF graph
-    :rtype: rdflib.Graph
     """
     import rdflib
 
@@ -765,15 +783,13 @@ def read_rdf(path: Union[str, Path], **kwargs):
         path = Path(path)
     graph = rdflib.Graph()
     with (
-        gzip.open(path, "rb")  # type: ignore
-        if isinstance(path, Path) and path.suffix == ".gz"
-        else open(path)
+        gzip.open(path, "rb") if isinstance(path, Path) and path.suffix == ".gz" else open(path)
     ) as file:
         graph.parse(file, **kwargs)
     return graph
 
 
-def write_sql(df, name: str, path: Union[str, Path], **kwargs) -> None:
+def write_sql(df: "pandas.DataFrame", name: str, path: Union[str, Path], **kwargs: Any) -> None:
     """Write a dataframe as a SQL table.
 
     :param df: A dataframe
@@ -874,7 +890,7 @@ def download_from_s3(
     s3_bucket: str,
     s3_key: str,
     path: Union[str, Path],
-    client=None,
+    client: Union[None, "botocore.client.BaseClient"] = None,
     client_kwargs: Optional[Mapping[str, Any]] = None,
     download_file_kwargs: Optional[Mapping[str, Any]] = None,
     force: bool = True,
diff --git a/tox.ini b/tox.ini
index 5b01431..0ff69bd 100644
--- a/tox.ini
+++ b/tox.ini
@@ -106,7 +106,7 @@ deps =
     types-requests
 skip_install = true
 commands =
-    mypy --install-types --non-interactive --ignore-missing-imports src/pystow/
+    mypy --install-types --non-interactive --ignore-missing-imports --strict src/pystow/
 description = Run the mypy tool to check static typing on the project.
 
 [testenv:pyroma]