diff --git a/pintxarray/__init__.py b/pintxarray/__init__.py
index 3e169c94..067a6a00 100644
--- a/pintxarray/__init__.py
+++ b/pintxarray/__init__.py
@@ -1 +1,2 @@
 from .accessors import PintDataArrayAccessor, PintDatasetAccessor
+from .indexing import QuantityIndex
diff --git a/pintxarray/accessors.py b/pintxarray/accessors.py
index 414efcda..e242f8c5 100644
--- a/pintxarray/accessors.py
+++ b/pintxarray/accessors.py
@@ -16,6 +16,7 @@
 if not IS_NEP18_ACTIVE:
     raise ImportError("NUMPY_EXPERIMENTAL_ARRAY_FUNCTION is not enabled")
 
+# TODO do we need this? xarray chooses it's own labels anyway
 try:
     mpl = import_module("matplotlib")
     has_mpl = True
diff --git a/pintxarray/indexing.py b/pintxarray/indexing.py
new file mode 100644
index 00000000..4edae7ba
--- /dev/null
+++ b/pintxarray/indexing.py
@@ -0,0 +1,29 @@
+import pandas as pd
+from pandas.core.base import IndexOpsMixin
+
+from pint import Quantity, Unit
+
+
+# TODO does pandas supply some kind of abstract base class for indexer objects?
+class QuantityIndex(IndexOpsMixin):
+    def __init__(self, index=pd.Index, units=None, unit_registry=None,
+                 decode_cf=False):
+        self._index = index
+        self._units = units
+
+    @property
+    def units(self):
+        ...
+
+    @property
+    def magnitude(self):
+        ...
+
+    def dequantify(self):
+        ...
+
+    def to(self, units):
+        ...
+
+    def to_base_units(self):
+        ...